02-Mass-Transit-Analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<script>\n",
       "  function code_toggle() {\n",
       "    if (code_shown){\n",
       "      $('div.input').hide('500');\n",
       "      $('#toggleButton').val('Show Code')\n",
       "    } else {\n",
       "      $('div.input').show('500');\n",
       "      $('#toggleButton').val('Hide Code')\n",
       "    }\n",
       "    code_shown = !code_shown\n",
       "  }\n",
       "\n",
       "  $( document ).ready(function(){\n",
       "    code_shown=false;\n",
       "    $('div.input').hide()\n",
       "  });\n",
       "</script>\n",
       "<form action=\"javascript:code_toggle()\"><input type=\"submit\" id=\"toggleButton\" value=\"Show Code\"></form>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import HTML\n",
    "HTML('''<script>\n",
    "  function code_toggle() {\n",
    "    if (code_shown){\n",
    "      $('div.input').hide('500');\n",
    "      $('#toggleButton').val('Show Code')\n",
    "    } else {\n",
    "      $('div.input').show('500');\n",
    "      $('#toggleButton').val('Hide Code')\n",
    "    }\n",
    "    code_shown = !code_shown\n",
    "  }\n",
    "\n",
    "  $( document ).ready(function(){\n",
    "    code_shown=false;\n",
    "    $('div.input').hide()\n",
    "  });\n",
    "</script>\n",
    "<form action=\"javascript:code_toggle()\"><input type=\"submit\" id=\"toggleButton\" value=\"Show Code\"></form>''')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Proposal: Develop Business for House Leasing Online Platform"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Abstract\n",
    "\n",
    "After second though, I decide to present the task as a pitch for the potential client, hope that works for this task. All the questions in the task will be answered in different sections below. So suppose we're now pitching to a client who does business on matching up rentor to landlord online. With the information gained from Shanghai transportation records, what can we do to further enhance and expand the client's business in Shanghai?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Picture an Application Scenario\n",
    "\n",
    "> Imagine our online flat leasing company would like to find out the most suitable subway stations to launch its ads, making most use of the budget and reaching the passengers who are most likely-to-be its customers. Furthermore, once we get the attention from them and direct them to our online platform, what additional information can we provide to further enhance their experience interacting with the platform?\n",
    "\n",
    "<mark>**Expected Outcome**</mark><br/>\n",
    "\n",
    "In order to answer the problem mentioned above, first I will identify the stations which have the highest share of long distance commuters, for these passengers would be most likely to be attracted by the flat ads at the station. Secondly, in their journey interacting with our platform, we can offer more flexible search methods for them to find out their favorite flat choices. One idea is to enable them to search based on the **expected commute time**. Lastly, the filter can be delved further to present the expected commute time over each hour within a day. This information will greatly benefit them in choosing appropriate flats, as they can now assess what' the best timing they should leave from home for office.\n",
    "\n",
    "<mark>**Actual Analysis Outcome**</mark>\n",
    "\n",
    "Let me quickly summarize the result in advance. By the end of this analysis, I am able to come up with a list of suggested stations for ads placement. Adhering  to the stations, I also obtain a table of expected commute time for various subway routes. Nonetheless, from the data exploration, I find there is no difference of commute time among hours on which passengers choose to take subway. But it's something await to be further investigated and confirmed. Noted that in this analysis, it will mostly be statistic summary, less of modeling, as I figure the statistics contain more business insights and values compared to modeling. Nonetheless, I do include a proposal for feature engineering and modeling under the condition if additional data are provided."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Structure of The Report\n",
    "\n",
    "<mark>**1. Data Processing**</mark><br/>\n",
    "\n",
    "In this section, practices include implementing sampling on original dataset, having data validation, and splitting out the subway data alone for later use.\n",
    "\n",
    "<mark>**2. Data Exploration on Subway Transportation Records**</mark><br/>\n",
    "\n",
    "I will find out what stations have the largest traffic volume, and what's the average commute time per subway trip.\n",
    "\n",
    "<mark>**3. Identify Best Stations for Ads Launch**</mark><br/>\n",
    "\n",
    "In this section, I will come up with a list of suggested stations for ads placement.\n",
    "\n",
    "<mark>**4. Get the Table of Expected Commute Time for each Route**</mark><br/>\n",
    "\n",
    "In the beginning, I will delve further to find commute time among different time slot. However, facing the limitation of fewer samples owing to sampling, I will just split the time slot into peak and off-peak hour. Then discuss if there shows difference of commute time between two time slots. If there is no difference of commute time between time slots. Then, I will just simplify the result dataset and compute the expected commute time for each route regardless of the timing to take subway.\n",
    "\n",
    "<mark>**5. Empower with more Additional Data**</mark><br/>\n",
    "\n",
    "In this section, I would discuss what are other suitable data to be collected and joined together. How these data will empower our flat leasing company even more? Let's say, our online flat leasing website already obtain each flat's rent, conditions etc, along with this traffic dataset. We can build up a rent prediction model out of these features.\n",
    "\n",
    "<mark>**6. Proposed Model to be Used and in What Way**</mark><br/>\n",
    "\n",
    "I will propose a hypothetical model that can predict the rent of the flat for newly landlord, based on the flat's location and other flat's conditions. This will be mostly helpful for the landlord, especially in some regions where there is little information of rent for reference. This section includes sketch of the model, what features to be considered and what models are appropriate in this case.\n",
    "\n",
    "<mark>**7. Other Applications from this Dataset**</mark><br/>\n",
    "\n",
    "In this final section, I will propose another application that this dataset is suitable to be used for. For instance, the traffic control of Shanghai Metro System.\n",
    "\n",
    "So, without further delay, let me begin today's exploration."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preprocess Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read in Raw Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import necessary modules\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import missingno as msno\n",
    "from tqdm import tqdm, tqdm_notebook\n",
    "\n",
    "import re\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "plt.rcParams['font.sans-serif'] = ['SimHei'] # 用来正常显示中文标签\n",
    "plt.rcParams['axes.unicode_minus'] = False # 用来正常显示负号\n",
    "sns.set_style('whitegrid',{'font.sans-serif':['SimHei','Arial']}) # 用来处理seaborn显示中文\n",
    "\n",
    "pd.options.display.float_format = '{:.2f}'.format\n",
    "\n",
    "from IPython.display import display"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to print bold text in print().\n",
    "from IPython.display import Markdown, display\n",
    "\n",
    "def printmd(string):\n",
    "    \"\"\"\n",
    "    Ref: https://discuss.analyticsvidhya.com/t/how-to-make-a-text-bold-within-print-statement-in-ipython-notebook/14552\n",
    "    \"\"\"\n",
    "    display(Markdown(string))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Read in dataset\n",
    "dataset = pd.read_csv('../Playground-dataset/02-Mass-Transit-Analysis-Dataset/SPTCC-20160701.csv', \n",
    "                      encoding='gbk',\n",
    "                      header=None,\n",
    "                      names = ['transit_card_id','date','time',\n",
    "                               'line_and_stop_name','method_of_transit','fare_collected','fare_type'],\n",
    "                      index_col=False) # encoding for simplified chinese\n",
    "\n",
    "# print('Data type of each column:\\n')\n",
    "# dataset.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Implement Sampling\n",
    "\n",
    "Since the original dataset is way too large, just for the efficiency of analysis later on, I will have it sampled and explore the result dataset afterwards. The sampling is conducted on **each passenger (from the transit_card_id), picking up one id out of every 25 ids.**\n",
    "\n",
    "<mark>**Result**</mark><br/>\n",
    "\n",
    "The original dataset contains 15,772,842 records, and the sampled dataset contains **630,163** recrods, roughly **96%** of the original dataset are dropped off in this analysis. This poses a problem when analysing the data in more granular level. Neverthelsee, it can be solved when re-run this anlaysis in larger scale and specify higher frquency of sampling."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Sort by card id and set it as index, for speeding up following sampling process.\n",
    "dataset = dataset.sort_values(by='transit_card_id').set_index(keys='transit_card_id')\n",
    "\n",
    "# Sampling based on transit_card_id.\n",
    "card_ids = np.unique(dataset.index.values)\n",
    "\n",
    "def sampling(df, ids, step=10):\n",
    "    \n",
    "    sub_df = pd.DataFrame()\n",
    "    for i in tqdm_notebook(range(0, len(ids), step)):\n",
    "        id_ = ids[i]\n",
    "        sub = df[df.index == id_]\n",
    "        sub_df = sub_df.append(sub)\n",
    "    \n",
    "    return sub_df\n",
    "\n",
    "samples = sampling(df=dataset, ids=card_ids, step=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the sampled dataset.\n",
    "pd.to_pickle(samples, '../Playground-dataset/02-Mass-Transit-Analysis-Dataset/samples.pickle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reload the dataset.\n",
    "samples = pd.read_pickle('../Playground-dataset/02-Mass-Transit-Analysis-Dataset/samples.pickle')\n",
    "samples = samples.reset_index()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check out Missing Value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**If exist missing values:**\n"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "transit_card_id       False\n",
      "date                  False\n",
      "time                  False\n",
      "line_and_stop_name    False\n",
      "method_of_transit     False\n",
      "fare_collected        False\n",
      "fare_type             False\n",
      "dtype: bool\n"
     ]
    }
   ],
   "source": [
    "printmd('**If exist missing values:**\\n')\n",
    "print(samples.isnull().any())\n",
    "\n",
    "# print('Show up the distribution of missing value over each column (marked as white lines for each record):')\n",
    "# _ = msno.matrix(samples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check out for Duplicates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**If exist duplicated rows:** True, **number of duplicated rows:** 1\n"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "**Rows after dropping off duplicates:** 630162"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "printmd('**If exist duplicated rows:** {}, **number of duplicated rows:** {}\\n'.format(\n",
    "    samples.duplicated().any(), samples.duplicated().sum()))\n",
    "\n",
    "samples = samples.drop_duplicates()\n",
    "printmd('**Rows after dropping off duplicates:** {}'.format(samples.shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Convert Datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "samples['datetime'] = samples.apply(lambda df: df['date'] + ' ' + df['time'], axis=1)\n",
    "samples['datetime'] = pd.to_datetime(samples['datetime'], format='%Y-%m-%d %H:%M:%S')\n",
    "\n",
    "samples = samples.drop(['date','time'], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Subset Dataset for Each Transportation Method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Unique value for method_of_transit:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['公交' '地铁' '出租' '轮渡' 'P+R停车场']\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "\n",
       "**Unique value for fare_type:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['非优惠' '优惠' '无']\n"
     ]
    }
   ],
   "source": [
    "# Preview of values in method_of_transit and fare_type.\n",
    "printmd('**Unique value for method_of_transit:**')\n",
    "print(samples['method_of_transit'].unique())\n",
    "\n",
    "printmd('\\n**Unique value for fare_type:**')\n",
    "print(samples['fare_type'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Separate dataset into each transportation type.\n",
    "subway = samples[samples['method_of_transit'] == '地铁'].sort_values(['transit_card_id','datetime']).reset_index(drop=True)\n",
    "ferry = samples[samples['method_of_transit'] == '轮渡'].sort_values(['transit_card_id','datetime']).reset_index(drop=True)\n",
    "taxi = samples[samples['method_of_transit'] == '出租'].sort_values(['transit_card_id','datetime']).reset_index(drop=True)\n",
    "parking = samples[samples['method_of_transit'] == 'P+R停车场'].sort_values(['transit_card_id','datetime']).reset_index(drop=True)\n",
    "bus = samples[samples['method_of_transit'] == '公交'].sort_values(['transit_card_id','datetime']).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Explore on Subway Trip Records"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Size of Subway:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(412196, 6)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "**Preview of Subway:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>transit_card_id</th>\n",
       "      <th>line_and_stop_name</th>\n",
       "      <th>method_of_transit</th>\n",
       "      <th>fare_collected</th>\n",
       "      <th>fare_type</th>\n",
       "      <th>datetime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>59911</td>\n",
       "      <td>2号线静安寺</td>\n",
       "      <td>地铁</td>\n",
       "      <td>0.00</td>\n",
       "      <td>非优惠</td>\n",
       "      <td>2016-07-01 20:10:57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>59911</td>\n",
       "      <td>8号线虹口足球场</td>\n",
       "      <td>地铁</td>\n",
       "      <td>4.00</td>\n",
       "      <td>非优惠</td>\n",
       "      <td>2016-07-01 20:36:52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>68421</td>\n",
       "      <td>8号线芦恒路</td>\n",
       "      <td>地铁</td>\n",
       "      <td>0.00</td>\n",
       "      <td>非优惠</td>\n",
       "      <td>2016-07-01 07:41:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>68421</td>\n",
       "      <td>9号线打浦桥</td>\n",
       "      <td>地铁</td>\n",
       "      <td>4.00</td>\n",
       "      <td>非优惠</td>\n",
       "      <td>2016-07-01 08:11:48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>68421</td>\n",
       "      <td>9号线打浦桥</td>\n",
       "      <td>地铁</td>\n",
       "      <td>0.00</td>\n",
       "      <td>非优惠</td>\n",
       "      <td>2016-07-01 15:26:27</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   transit_card_id line_and_stop_name method_of_transit  fare_collected  \\\n",
       "0            59911             2号线静安寺                地铁            0.00   \n",
       "1            59911           8号线虹口足球场                地铁            4.00   \n",
       "2            68421             8号线芦恒路                地铁            0.00   \n",
       "3            68421             9号线打浦桥                地铁            4.00   \n",
       "4            68421             9号线打浦桥                地铁            0.00   \n",
       "\n",
       "  fare_type            datetime  \n",
       "0       非优惠 2016-07-01 20:10:57  \n",
       "1       非优惠 2016-07-01 20:36:52  \n",
       "2       非优惠 2016-07-01 07:41:04  \n",
       "3       非优惠 2016-07-01 08:11:48  \n",
       "4       非优惠 2016-07-01 15:26:27  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "printmd('**Size of Subway:**')\n",
    "display(subway.shape)\n",
    "\n",
    "printmd('**Preview of Subway:**')\n",
    "display(subway.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clean up and Validate Quality in Data\n",
    "\n",
    "As we saw from dataset above, every passenger will **leave two records** - entry and exit of subway station for one trip. Thus, I will implement the following steps to further ensure data quality and transform subway dataset into ready-for-use, clean data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<mark>**Goal**</mark><br/>\n",
    "\n",
    "Ensure each id has one *entry* record and corresponding one *exit* record. It will be checked by my self-defined functions `criterion()` and `df_concat()`.\n",
    "\n",
    "<span style='margin:15px'>\n",
    "    \n",
    "1. I assume the entry record occurs in odd rows and has value 0 in *fare_collected* column over each id.\n",
    "2. The exit record must follow immediately the entry record regardless what value it has in fare_collected column. The reasoning behind is that final fares can change due to various reasons, promotion, cash rewards etc. Thus, no need to examine the fare being collected.\n",
    "3. Record that doesn't comply with the rules above will be removed from dataset.\n",
    "4. The clean dataset for each id must have even rows (entry and exit pair-up) and eventually, join data of all id together as our final analysis dataset.\n",
    "5. Split the *line* and *station name* from single column to separate columns.\n",
    "\n",
    "</span>\n",
    "\n",
    "\n",
    "<mark>**Result**</mark><br/>\n",
    "\n",
    "- `Subway_clean` is ready for use in analysing traffic volume for each station.\n",
    "- **8278** rows are being dropped off from subway dataset owing to the process above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# Drop off weird rows.\n",
    "# criterion: The value of fare_collected in even index must be 0. \n",
    "#       if not, then remove the row and start examining the df all over again -> python is zero-based numbering\n",
    "# df_concat: Only append df to the final dataset if df's size is even (2,4,6...etc) excluding 0.\n",
    "#       if not, then drop off the last row and append remainging to final dataset.\n",
    "\n",
    "\n",
    "def criterion(df):\n",
    "    \"\"\" \n",
    "    @return\n",
    "        df: can be shape of 0, 1, 2, 3, etc.\n",
    "    \"\"\"\n",
    "    df.reset_index(drop=True, inplace=True)\n",
    "    idx = 0\n",
    "    max_idx = df.index.max()\n",
    "    while True:\n",
    "        if idx % 2 == 0: # for even index\n",
    "            if df['fare_collected'][idx] != 0:\n",
    "                df = df.drop(df.index[idx]) # remove the record\n",
    "                df.reset_index(drop=True, inplace=True) # reset index\n",
    "                if df.shape[0] == 0:\n",
    "                    break # if df is null, then exit the loop\n",
    "                else:\n",
    "                    idx = 0 # loop from the beginning again\n",
    "                    max_idx = df.index.max()\n",
    "            else:\n",
    "                idx += 1\n",
    "        else: # for odd index\n",
    "            idx += 1\n",
    "        \n",
    "        if idx >= max_idx:\n",
    "            break\n",
    "    return df\n",
    "    \n",
    "def df_concat(final_df, df):\n",
    "    \"\"\"\n",
    "    @param\n",
    "        final_df: final clean dataset\n",
    "        df: the dataset to be appended to final_df\n",
    "    @return\n",
    "        final_df\n",
    "    \"\"\"\n",
    "    \n",
    "    if df.shape[0] % 2 == 0: # if df shape is 0 or even\n",
    "        if df.shape[0] != 0: # if df shape is not 0\n",
    "            final_df = final_df.append(df)\n",
    "    else:\n",
    "        if df.shape[0] != 1: # if df shape is not 1\n",
    "            df = df.iloc[:-1, :] # drop off the last row\n",
    "            final_df = final_df.append(df)\n",
    "    return final_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# prepare the process status bar\n",
    "iters = range(len(subway['transit_card_id'].unique()))\n",
    "\n",
    "subway_clean = pd.DataFrame()\n",
    "\n",
    "with tqdm_notebook(total=len(iters)) as pbar:\n",
    "    for k1, group in subway.groupby(['transit_card_id']):\n",
    "        group = criterion(group)\n",
    "        subway_clean = df_concat(subway_clean, group)\n",
    "        pbar.update(1)\n",
    "    \n",
    "# reset index\n",
    "subway_clean.reset_index(drop=True, inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split line number and station name\n",
    "subway_clean['line_and_stop_name'] = subway_clean['line_and_stop_name'].str.strip() # stip string first\n",
    "regex = re.compile('(\\d+)号线(\\w+)')\n",
    "col = subway_clean['line_and_stop_name'].apply(lambda x: regex.search(x).group(1,2))\n",
    "line, stop = map(np.array, zip(*col))\n",
    "subway_clean['line'] = line\n",
    "subway_clean['stop'] = stop\n",
    "subway_clean.drop(labels=['line_and_stop_name'], axis=1, inplace=True)\n",
    "print('Preview of subway_clean:')\n",
    "display(subway_clean.head(n=2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preview of subway_clean\n",
    "printmd('**Size of subway_clean:**')\n",
    "display(subway_clean.shape)\n",
    "printmd('**Preview of subway_clean:**')\n",
    "display(subway_clean.head(n=8))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# backup file\n",
    "pd.to_pickle(subway_clean, '../Playground-dataset/02-Mass-Transit-Analysis-Dataset/subway_clean.pickle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "subway_clean = pd.read_pickle('../Playground-dataset/02-Mass-Transit-Analysis-Dataset/subway_clean.pickle')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Stations with Largest Traffic\n",
    "\n",
    "First, we can tell from the table below, the top stations with largest traffic volumn taking both entries and exits into account are:\n",
    "\n",
    "<mark>**NOTED**</mark><br/>\n",
    "This result comes from sampled dataset. The absolute values are under-estimated."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "subway_clean = pd.read_pickle('../Playground-dataset/02-Mass-Transit-Analysis-Dataset/subway_clean.pickle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "traffic_volumn = subway_clean.groupby(['line', 'stop']).count(\n",
    "    )[['transit_card_id']].rename(columns={'transit_card_id':'traffic_volumn'}\n",
    "                                 ).sort_values(by=['traffic_volumn'], ascending=False)\n",
    "\n",
    "print('First 5 stations with largest traffic:')\n",
    "display(traffic_volumn.head(n=5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot station with largest traffic.\n",
    "fig, ax = plt.subplots(1,1,figsize=(15,5))\n",
    "\n",
    "_ = traffic_volumn.iloc[:10,:].plot.bar(y=['traffic_volumn'], ax=ax, rot=30, colors='#cec3c6')\n",
    "_ = ax.legend(fontsize='x-large')\n",
    "_ = ax.set_title('Stations of Largest Traffic', fontsize=20)\n",
    "_ = ax.set_xlabel(xlabel='Line,Stations', fontsize=15)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Average Subway Trip Length"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prepare Dataset\n",
    "\n",
    "Before compute the average trip length, let's transform the data again and create a new dataset `subway_trip` to record data in format of **per trip per row** for each id.\n",
    "\n",
    "More specifically, each row will contain four columns: get-on station, time of get-on, get-off station, and time of get-off etc."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# prepare dataset for trip length\n",
    "subway_trip_start = subway_clean[::2].reset_index(drop=True) # even row\n",
    "subway_trip_start.columns = [col + '_x' for col in subway_trip_start.columns.values]\n",
    "subway_trip_end = subway_clean[1::2].reset_index(drop=True) # odd row\n",
    "subway_trip_end.columns = [col + '_y' for col in subway_trip_end.columns.values]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Preview of subway_trip:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>transit_card_id_x</th>\n",
       "      <th>line_x</th>\n",
       "      <th>stop_x</th>\n",
       "      <th>datetime_x</th>\n",
       "      <th>line_y</th>\n",
       "      <th>stop_y</th>\n",
       "      <th>datetime_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>59911</td>\n",
       "      <td>2</td>\n",
       "      <td>静安寺</td>\n",
       "      <td>2016-07-01 20:10:57</td>\n",
       "      <td>8</td>\n",
       "      <td>虹口足球场</td>\n",
       "      <td>2016-07-01 20:36:52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>68421</td>\n",
       "      <td>8</td>\n",
       "      <td>芦恒路</td>\n",
       "      <td>2016-07-01 07:41:04</td>\n",
       "      <td>9</td>\n",
       "      <td>打浦桥</td>\n",
       "      <td>2016-07-01 08:11:48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>68421</td>\n",
       "      <td>9</td>\n",
       "      <td>打浦桥</td>\n",
       "      <td>2016-07-01 15:26:27</td>\n",
       "      <td>8</td>\n",
       "      <td>芦恒路</td>\n",
       "      <td>2016-07-01 16:11:32</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   transit_card_id_x line_x stop_x          datetime_x line_y stop_y  \\\n",
       "0              59911      2    静安寺 2016-07-01 20:10:57      8  虹口足球场   \n",
       "1              68421      8    芦恒路 2016-07-01 07:41:04      9    打浦桥   \n",
       "2              68421      9    打浦桥 2016-07-01 15:26:27      8    芦恒路   \n",
       "\n",
       "           datetime_y  \n",
       "0 2016-07-01 20:36:52  \n",
       "1 2016-07-01 08:11:48  \n",
       "2 2016-07-01 16:11:32  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Concatnate two dataset.\n",
    "subway_trip = pd.concat([subway_trip_start, subway_trip_end], axis=1)\n",
    "printmd('**Preview of subway_trip:**')\n",
    "display(subway_trip[['transit_card_id_x','line_x','stop_x','datetime_x',\n",
    "                    'line_y','stop_y','datetime_y']].head(n=3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save backup file.\n",
    "pd.to_pickle(subway_trip, '../Playground-dataset/02-Mass-Transit-Analysis-Dataset/subway_trip.pickle')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find out Average Trip Length\n",
    "\n",
    "<mark>**Result**</mark><br/>\n",
    "\n",
    "We can tell that average subway tip length is roughly **43 mins**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "subway_trip = pd.read_pickle('../Playground-dataset/02-Mass-Transit-Analysis-Dataset/subway_trip.pickle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Preview of trip_length:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>transit_card_id_x</th>\n",
       "      <th>line_x</th>\n",
       "      <th>stop_x</th>\n",
       "      <th>datetime_x</th>\n",
       "      <th>line_y</th>\n",
       "      <th>stop_y</th>\n",
       "      <th>datetime_y</th>\n",
       "      <th>trip_length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>59911</td>\n",
       "      <td>2</td>\n",
       "      <td>静安寺</td>\n",
       "      <td>2016-07-01 20:10:57</td>\n",
       "      <td>8</td>\n",
       "      <td>虹口足球场</td>\n",
       "      <td>2016-07-01 20:36:52</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>68421</td>\n",
       "      <td>8</td>\n",
       "      <td>芦恒路</td>\n",
       "      <td>2016-07-01 07:41:04</td>\n",
       "      <td>9</td>\n",
       "      <td>打浦桥</td>\n",
       "      <td>2016-07-01 08:11:48</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>68421</td>\n",
       "      <td>9</td>\n",
       "      <td>打浦桥</td>\n",
       "      <td>2016-07-01 15:26:27</td>\n",
       "      <td>8</td>\n",
       "      <td>芦恒路</td>\n",
       "      <td>2016-07-01 16:11:32</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>80212</td>\n",
       "      <td>4</td>\n",
       "      <td>南浦大桥</td>\n",
       "      <td>2016-07-01 18:22:12</td>\n",
       "      <td>4</td>\n",
       "      <td>宜山路</td>\n",
       "      <td>2016-07-01 18:42:30</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>85716</td>\n",
       "      <td>6</td>\n",
       "      <td>外高桥保税区南</td>\n",
       "      <td>2016-07-01 13:12:39</td>\n",
       "      <td>6</td>\n",
       "      <td>洲海路</td>\n",
       "      <td>2016-07-01 13:17:08</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   transit_card_id_x line_x   stop_x          datetime_x line_y stop_y  \\\n",
       "0              59911      2      静安寺 2016-07-01 20:10:57      8  虹口足球场   \n",
       "1              68421      8      芦恒路 2016-07-01 07:41:04      9    打浦桥   \n",
       "2              68421      9      打浦桥 2016-07-01 15:26:27      8    芦恒路   \n",
       "3              80212      4     南浦大桥 2016-07-01 18:22:12      4    宜山路   \n",
       "4              85716      6  外高桥保税区南 2016-07-01 13:12:39      6    洲海路   \n",
       "\n",
       "           datetime_y  trip_length  \n",
       "0 2016-07-01 20:36:52           25  \n",
       "1 2016-07-01 08:11:48           30  \n",
       "2 2016-07-01 16:11:32           45  \n",
       "3 2016-07-01 18:42:30           20  \n",
       "4 2016-07-01 13:17:08            4  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/markdown": [
       "**Mean Trip Length:** 42.94 mins"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Compute trip length.\n",
    "subway_trip['trip_length'] = subway_trip['datetime_y'] - subway_trip['datetime_x']\n",
    "\n",
    "# Convert trip length to minutes integers.\n",
    "subway_trip['trip_length'] = subway_trip['trip_length'].apply(lambda x: int(x.total_seconds()//60))\n",
    "\n",
    "printmd('**Preview of trip_length:**')\n",
    "display(subway_trip[['transit_card_id_x','line_x','stop_x','datetime_x',\n",
    "                     'line_y','stop_y','datetime_y','trip_length']].head(n=5))\n",
    "printmd('**Mean Trip Length:** {:.2f} mins'.format(subway_trip['trip_length'].mean()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Identify Best Stations to Place Rent Ads\n",
    "\n",
    "It's time to use the processed dataset to help us make practical decisions on marketing - outdoor ads launch in this case.\n",
    "\n",
    "Over all 413 stations in Shanghai Metro, you'd probably want to know **which station has the highest share of long distance passengers**. For it will be most effective to place outdoor ads for flat leasing at these stations, for long distance passengers are definitely in need of flats of best price with least commute time."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Distribution of Trip Length in The Morning\n",
    "\n",
    "In the beginning, we need to filter on subway trip occurred in the morning (remaining 80,007 records). I just want to target the **morning commuters**, since they're most likely to be interested in finding new flats closer to the office. Thus, what's distribution like for morning trip length?\n",
    "\n",
    "Form data shown on the graph below, trip length is extremely right-skewed, where the mean is **43 mins** and the median is **32 mins**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Subway trips occurred in the moring:  80007\n"
     ]
    }
   ],
   "source": [
    "# Subset trip occurs before 12pm.\n",
    "subway_trip_mor = subway_trip[subway_trip['datetime_y'].apply(lambda x: x.hour < 12)]\n",
    "print('Subway trips occurred in the moring: ', subway_trip_mor.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_distribution(df, col, bins=15):\n",
    "    \n",
    "    fig, ax = plt.subplots(figsize=(15,5))\n",
    "    sns.set_style(style='whitegrid')\n",
    "    feature = df[col]\n",
    "    sns.distplot(feature, bins=bins, hist=True, kde=True, ax=ax, color='#cec3c6')\n",
    "    mean_line = ax.axvline(x=feature.mean(), linewidth=2, color='black', label='mean {:.2f}'.format(feature.mean()))\n",
    "    median_line = ax.axvline(x=feature.median(), linewidth=2, color='red', label='median {:.2f}'.format(feature.median()))\n",
    "    _ = ax.legend(handles=[mean_line, median_line])\n",
    "    _ = ax.set_title('Histogram of {} for Bins = {}'.format(col, bins), fontsize='x-large')\n",
    "    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Descriptive Statistics of Trip Length:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>trip_length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>80007.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>35.69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>21.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>21.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>32.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>47.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>362.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       trip_length\n",
       "count     80007.00\n",
       "mean         35.69\n",
       "std          21.08\n",
       "min           0.00\n",
       "25%          21.00\n",
       "50%          32.00\n",
       "75%          47.00\n",
       "max         362.00"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "printmd('**Descriptive Statistics of Trip Length:**')\n",
    "display(subway_trip_mor[['trip_length']].describe())\n",
    "\n",
    "plot_distribution(df=subway_trip_mor, col='trip_length', bins=25)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## After Delete Outliers\n",
    "\n",
    "It's obvious there are some outliers significantly dragging the distribution away, thus negatively influencing our interpretation. Let's remove them and observe the refined result, see if it improves.\n",
    "\n",
    "After removing the outliers (3854 records deleted), we see the distribution is less right-skewed. Besides, from the histogram below, we know that most passengers' trip length span from **20 to 40 mins** (accounts more than half of all trips)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Outliers being deleted: 3854\n"
     ]
    }
   ],
   "source": [
    "# Find outliers using IQR.\n",
    "Q3 = float(subway_trip_mor['trip_length'].quantile([0.75]))\n",
    "Q1 = float(subway_trip_mor['trip_length'].quantile([0.25]))\n",
    "IQR = Q3 - Q1\n",
    "\n",
    "high = Q3 + IQR\n",
    "low = Q1 - IQR\n",
    "\n",
    "subway_trip_mor_clean = subway_trip_mor[(low <= subway_trip_mor['trip_length']) & (subway_trip_mor['trip_length'] <= high)]\n",
    "print('Outliers being deleted: {}'.format(subway_trip_mor.shape[0] - subway_trip_mor_clean.shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Descriptive Statistics of Trip Length after Outliers Removed:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>trip_length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>76153.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>32.87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>16.31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>20.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>31.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>44.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>73.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       trip_length\n",
       "count     76153.00\n",
       "mean         32.87\n",
       "std          16.31\n",
       "min           0.00\n",
       "25%          20.00\n",
       "50%          31.00\n",
       "75%          44.00\n",
       "max          73.00"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3YAAAFMCAYAAACH0y5vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Xl8VPW9//HX7DOZmUz2nbApiIqy2Fq3y4VSW6UqxbrxI9ba1qVab20RuIrWpSqo1VYqqFetgrWVKna5Lm1tqVZv5V6xiCIIhi17JntmMpnMcn5/BKYEAgkkYZLwfj7MI8ycOef7+Z45zpxPvpvJMAwDERERERERGbLMyQ5ARERERERE+kaJnYiIiIiIyBCnxE5ERERERGSIU2InIiIiIiIyxCmxExERERERGeKU2ImIiIiIiAxxSuxERPrJjBkzeP755w94vry8nPHjx7N161YAFi1axE033dSrY/75z3+mqqqqX+McbB5++GGmTJnC1KlTaW5uPmB7T+dg2bJlzJkzp19iWbduHePHjycYDPbL8XpSXl7Om2++mXh8sGvoYPx+P5dffjkTJ05k/vz5/RLTjBkzGD9+fOJn8uTJzJs3j08++STxmjVr1nD66af3S3l9FQ6HeeCBB/j3f/93pk6dytVXX01paWlie01NTZf67P1paGhIYtQiIv3PmuwARESONbfddhu9WUK0oqKCG2+8kT/84Q/k5+cfhciOvpqaGp544gnuuecezjzzTHw+X5ftvTkHV199NfPmzTsa4fa7//zP/+Tkk09m5syZR7T/Sy+9RHl5Ob/97W9JS0vrt7h+8IMfMGfOHAzDoKWlhZUrV3LNNdfw5z//GZfLxfnnn8+0adP6rby+eOihh1i7di1LliwhJyeHFStWcPXVV/Pqq6/i8XgoLS0lNTWV1157rct+6enpSYpYRGRgqMVOROQo83q9pKam9vi63iR/Q11LSwsAZ5xxBkVFRQds7805cLvdx+xNektLC6NGjWLs2LFkZmb223HdbjfZ2dnk5ORw3HHHccstt1BXV8c//vEPAJxOZ7+Wd6Ti8Thr1qzhBz/4AV/4whcYM2YM9957L83Nzbz33nsAbNu2jTFjxpCdnd3lx2QyJTl6EZH+pcROROQo27crZiAQ4Ic//CGnn346kyZN4lvf+hY7d+4E4Itf/CIAF1xwAcuWLQPgo48+4sorr2Ty5MmcffbZPPjgg0QikcSx//GPf3DRRRdxyimnMHfuXB599FFKSkqAzu5zX/va15g/fz5Tpkxh+fLlRKNRHnnkEWbMmMFJJ53EmWeeyY9//GNisRjQ2c3xe9/7Hg899BBTp07lC1/4AitXruT999/nggsuYNKkSXz729+mqanpoPV95513uOSSS5g0aRLTp0/nqaeewjAM1q1bx1e/+lUAZs6cyaJFiw7Yd/9z0F0d9u2KuW7dOk477TTWrFnD2WefzWmnncbtt99Oe3v7Eb1XgUCA22+/nc9//vOcfvrp3HTTTdTU1CS2z5gxg+eee46SkhJOPfVULrjgAv76178mtjc3N/P973+fKVOmMG3aNNasWcOJJ55IeXk5ixYt4n//93955plnmDFjRmKf3bt3841vfIOJEycyc+ZMXn/99W5jW7RoEc888wz/93//x/jx41m3bh2GYbBy5Uq+/OUvM3HiRC666CLeeuutxD4lJSXceeednHfeeZxxxhlduiweitvtxm63Jx7v2xVzb1fjN954g/POO4/JkydTUlLC9u3bE69/7LHHmDZtGhMnTmT27NldYtrX3q6w3f3s/X9gX/F4nEcffZQzzzwz8ZzZ3HlrEw6HAfjss88YM2ZMr+opIjKUKbETEUmin/70p5SXl7Ny5UrWrFmD2Wzm1ltvBeA3v/kNAKtWreLqq69mx44dlJSUcNxxx/HSSy9xzz338Lvf/Y6HH34YgLKyMq699lqmTZvGb3/7W7785S/zxBNPdCnvk08+we1288orrzB79myefvppfve737FkyRL++Mc/smDBAl544YUu477Wrl1LW1sbr7zyCpdffjlLlizhnnvu4Uc/+hFPP/00mzZt4rnnnuu2fv/3f//HNddcw/Tp03nllVe4+eabWb58OS+88AKTJ09m1apVibredtttB+y//znorg77a2tr4+mnn+bRRx9lxYoV/M///A933XXXYb0ve91xxx3s2LGDp556ilWrVmEymfj2t79NNBpNvObRRx/l8ssv5+WXX2bUqFH853/+Jx0dHUBnl8aysjJWrlzJgw8+yOOPP55Imm+77TYmT57MFVdcwUsvvZQ43urVq/n617/Oq6++yplnnsmCBQsIBAIHxHbbbbdxxRVXMHnyZN555x0mT57M448/zrJly7jpppv4/e9/z8yZM7n++uvZsmVLYr+XXnqJW2+9lSeeeIKxY8f2eA46Ojp46qmnyMjI4Iwzzjjo637+859z991389xzz+H3+3nggQeAzjGSe+v/+uuvM23aNL7//e93W6e9denuZ+/7vy+r1cpZZ53VpRvqiy++SEdHB5MmTQI6W+xqamq45JJLOPvss7nuuuvYtWtXj/UWERlqNMZORKQfLVmyhJ/85CddnjtUd8KKigrcbjdFRUW43W5+/OMfU1FRAUBGRgYAaWlpuN1uVq9ezYgRI7j99tsxmUyMHTuWW2+9lQULFvC9732P3/zmN4wdO5Yf/OAHAIwZM4YNGzZQV1fXpcwbb7yR7OxsAI477jjuv/9+Pv/5zwNQVFTE008/zWeffcaXv/xlABwOB7feeitWq5WSkhJWrFjBvHnzOO200wCYNm0an332Wbf1W7VqFeeccw7f/e53ARg9ejTV1dU8/vjj/L//9/8SN+QZGRl4vd4D9t//HHRXh/3FYjHuuusupkyZAnSOY/v+97/PbbfdhsfjOeh7sb+ysjJeffVV3n77bXJzcwF48MEHOf300/n73//O9OnTAZg1axazZs0C4IYbbuBPf/pT4j185513+N3vfscJJ5wAwOLFi/nOd74DdHbJtdlsuFyuRD0Bvv71r3PBBRcA8N3vfpcXX3yR7du3c8opp3SJz+v14nK5sNlsZGdnYxgGzz33HNddd10inu9973t8+OGH/Nd//Vfiujz99NM555xzDln3fa/j9vZ24vE4d911Fy6X66D7XH/99Xzuc58DYO7cuTz55JNA5zVus9koKCigqKiIG264gc997nNYrQfegtjt9oO+r73xv//7vyxdupRvfvObFBYWArB9+3ZGjx7NwoULsVqtPP7448ybN49XX321V12iRUSGCiV2IiL96Nprr+XCCy/s8lxNTU2iO2R3r7/22ms544wz+NznPsfMmTO56KKLun3ttm3bOPXUU7uMDZo6dSqRSIRdu3bx6aefMnHixC77TJo0qUvrm8vl6nLj/MUvfpH33nuPBx54gJ07d/Lpp59SUVHBueeem3hNYWFh4ibc6XQCdBkPZ7fbaWxsPGjM+5+PqVOn8pOf/CQxvu5w7V+H/VkslkRrDcDEiROJRCLdJkeHsjdZ/cpXvtLl+VAoxI4dOxKJ3ejRoxPb9iaO0WiU0tJS7HY748ePT2yfPHlyj+WOGDEi8e+9yW5vupLW19fT2NjYpe7Qeb7feOONxOPi4uIej7XvdRwKhVi/fj333XcfTqez21ZSgFGjRiX+7fF4Eq2aF154Ia+88gozZ87kpJNOYsaMGVx88cWJa2lf77//fiLx7S6m66677qAxv/3229x0001Mnz6dH/7wh4nn33rrLSwWS6Ir6U9/+lP+/d//nT/+8Y9ccsklhz4RIiJDiBI7EZF+lJ6ezsiRI7s8Z7FYDvr6SZMm8Ze//IW1a9fy9ttv88gjj/DCCy906Zq3V3c3wntbA+PxOFarlXg8fsj4HA5Hl8fLli1j5cqVXHzxxZx77rnMnz8/0eK3V3ctK3vHMfWkp5iPxP512J/ZbO4S397yDvU+dCcWi2Gz2XjllVcOmGhj39k7bTbbAfsahoHVaj2iCXC6i7M3x+nuXO/dd99zfbDX7Wv/6/iEE07g008/ZeXKlQdN7PY/D3tjzsjIYM2aNfzjH//gb3/7G6+88grPPfcczz//POPGjeuyz8knn8xvf/vbbo+//4yp+3rttddYsGAB5513HkuWLOny/u/fyuh0OikqKqK6uvqgxxMRGYo0xk5EJImeeOIJNm7cyAUXXMCDDz7Iiy++yNatW/n0008PSCbGjh3Lhx9+2OUm/4MPPsBms1FcXMzxxx/Ppk2buuzz0UcfHbL8X/7ylyxcuJBFixYxe/ZsioqKqKys7LcZOfd2B93XP//5TzIzMw95o77XkcxcGIlEEmsGAmzcuBGHw9GlZa03xowZQyQSIRQKMXLkSEaOHEl2dnaidbMnxx9/PJFIhE8//TTxXE/vR194PB5ycnK6Pd/9MXlIPB4/ouviT3/6E7/+9a85++yzWbx4MW+88QZut7vbCVScTmfiXO//c7DlHN555x1uueUW5syZwwMPPNAlMa6pqWHq1Km8//77iecCgQA7d+7s1fhCEZGhRImdiEgSVVdX8+Mf/5gPPviAsrIyXnnlFTweD6NGjSIlJQWAzZs309rayty5cykvL+eee+6htLSUt956i6VLlzJ79my8Xi9XXHEFpaWlPPLII+zYsYNf/epXB51Rca+0tDTeeustdu3axaZNm7jppptobm5OTP7RV9/+9rf5+9//zvLly9m5cyevvfYaTz75JCUlJb1K2vY/B711++23s2nTJtatW8eSJUu45JJLEsfqrTFjxjBjxgwWLFjA+++/T2lpKQsXLuTDDz/sVaI0cuRIpk+fzu23385HH33EBx98wD333AP8K2F1u93s2rWry0ybfXHNNdfw+OOP8+qrr7Jz506WL1/OO++8c9CuwAcTDAbx+/34/X6qq6v5/e9/zx/+8IfE2L3DEYvFePDBB/njH/9IRUUFf/rTn/D7/Zx88smHfaz9hcNhFi1axOTJk7nxxhupq6tLxB0KhcjNzeXEE0/kxz/+MR9++CFbtmzhhz/8IVlZWXzpS1/qc/kiIoOJumKKiCTRLbfcwn333ceNN95Ia2srEyZM4Mknn0xM6vD1r3+dxYsXc/nll3Pbbbfx1FNP8eCDD3LRRReRnp7OnDlzuOGGGwDIzc1l+fLl3H///Tz99NNMmjSJCy+8kNra2oOWv2TJEu666y4uuOACMjIymDlzJl//+tcPaPk7UhMmTGDZsmX87Gc/Y/ny5eTl5fHd736Xq666qlf7p6endzkHEyZM6NV+s2bN4lvf+haGYTBnzpwDupf21tKlS7n//vu54YYb6Ojo4NRTT+WZZ57p9aQb9913H3fccQfz5s3D5/Mxb948fvKTnyS6LV5xxRUsXLiQCy+8MLFGXF/MmzePtrY2HnzwQerr6xk3bhyPP/54YqKb3nr44YcTs63abDYKCwv53ve+1+3MlD0577zzqKys5IEHHqCmpoaCggLuuOOOQ86w2Vvr169PJHL7Twhz++23M2/ePH72s5+xdOlSrrvuOkKhEGeccQZPP/10t11oRUSGMpNxLKyAKyJyDNi6dSvt7e1dJgi56667aG9v5/77709iZEfPunXruPLKK/nggw+6zKKZDKFQiHfffZd/+7d/S0zcsXHjRubOncuGDRu6HbsoIiJypNQVU0RkmCgvL+eqq67ib3/7GxUVFbzxxhv89re/5fzzz092aMckh8PB4sWLefjhhykrK2PTpk0sXbqUL3/5y0rqRESk3+mbRURkmJgxYwbXXXcdd999N36/n6KiIu64444e1yw7Vvzxj39k0aJFh3zNu+++e9hj8Q7GbDazYsUKli5dyq9//WucTifnnnsuCxYs6Jfji4iI7EtdMUVE5JgQDAYPWKx9f8XFxUc0E6eIiEiyKbETEREREREZ4jTGTkREREREZIgb1GPsNmzYgMPhSHYYw1Y4HNb5lWFF1/QgsndB6MOcZr/rId7fc4gjP8ZQp2tahiNd1zLcDPQ1HQ6HmTRpUo+vG9SJncPh6PWaRXL4Nm/erPMrw4qu6UHkxBM7f/eht/+Je45xLI8Y0DUtw5GuaxluBvqa3rx5c69ep66YIiIiIiIiQ5wSOxERERERkSFOiZ2IiIiIiMgQN6jH2ImIiIiIyNERiUQoLy+nvb092aEMKZFIpNfj4A7F6XRSVFSEzWY7ov2V2ImIiIiICOXl5Xi9XkaNGoXJZEp2OENGKBTC5XL16RiGYVBfX095eTmjR48+omOoK6aIiIiIiNDe3k5mZqaSuiQwmUxkZmb2qbVUiZ2IiIiIiAAoqUuivp57dcUUEREREZFjTltbGz/84Q9pbm7G5XLx4IMPkpGRwT/+8Q9++tOfYrVayczMZOnSpV26Wra2tnLzzTcTCoWw2Wzcc889jBgxgpKSksRrtm/fzte+9jXmz59/1OqjFjsRERERETnmrF69mpNOOokXXniBWbNmsXz5cgDuvPNOHnvsMX75y18ycuRIfvOb33TZb82aNYwbN45f/vKXnH/++Tz33HMArFq1ilWrVnHfffeRm5vL9ddff1TroxY7ERERERFJujVr1rB27Vra29vx+/1ceeWV/OUvf2Hbtm0sWLCAmTNn8vrrr/Pss89iNpuZOnUq8+fPp7q6mjvvvJNwOExTUxM33HADM2fO5IILLuDzn/88n376KSaTieXLl+P1ehPlXXXVVcRiMQAqKyvJysoCOhO0vf+ORqM4HI4ucY4bN47t27cDEAgEsFq7plT33nsvt9xyC263e8DOVXfUYiciIiIiIl3MmjULk8nUrz+zZs3qsdxgMMh//dd/8Z3vfIdf/epX/PznP+fuu+9mzZo1NDU1sWzZMp599ll+9atfUVNTw7vvvsv27dv55je/yS9+8Qtuv/12fvnLXyaONWvWLJ5//nlycnJ4++23DyjPYrFw5ZVX8vzzzzNt2jQAcnJyAPjzn//MunXrmD17dpd90tPTeffddzn//PN5+umnu2zfsmULwWCQM84444jP/ZFSi52ISB9FI1FisehRKctisWK1DfxHd1/rtPdvm+FezO51tOokIiKD34QJEwDwer2MHTsWk8mEz+cjHA6ze/duGhoauOaaa4DOxK2srIypU6eyYsUKXnrpJUwmE9Hov76/TjzxRADy8/MJh8Pdlrly5UpKS0u59tprefPNNwF49tlneeONN3jqqacOaLH7+c9/zre//W0uv/xytmzZwvz58/nv//5vAH7/+99zySWX9O9J6SV9k4qI9FEsFqW1qfmolOVN8x2VJKivddr7FdibYxytOomISO+9+uqrSSn3UDNDFhUVkZ+fzzPPPIPNZmPNmjVMmDCBn/3sZ1xyySVMmzaNl19+mVdeeaVXx3viiSfIzc1l9uzZpKSkYLFYAFixYgWbNm3i2Wefxel0HrBfampqoktnZmYmgUAgse29997jO9/5zmHXuz/om1RERERERAa9jIwMrrrqKkpKSojFYhQWFnLeeefxla98hXvvvZcnnniC/Px8Ghsbe3W8iy++mIULF/Lyyy8Ti8W47777qKur47HHHuPEE09MJGjnnXcec+fO5eqrr+bxxx/nP/7jP1i8eDEvvPAC0WiUO+64I3FMv99Penr6gNS/JybDMIyklNwLmzdvTjTHSv/T+ZXhJlnXdLi9/ai22Dm6+ethf+trnbLy8wCoq6ru8bUHq9Pev7IO4q+pAafPaRmOdF0PXnpvjkwoFOqyHEJfdPce9PZ9UYudiMgQYsSNXo1b63s58QEvQ0RERPqPEjsRkSEkHo/T1hLo+YV9lOLxDHgZIiIi0n+03IGIiIiIiMgQpxY7EZFBJh6PE4tGiUVjxGMxYrFY5+NYjKa6eqKRKGDQOfSs87fJZMJkNmE2mTCZzZhM5s7HZlPi3yaTOfHYbDZjsuz5vWd9IRERERm6lNiJiCSJYRjEolE6wh1EwmEiHR10hDuIRbtfP85itWC2WDAMAxMmOv/rTMjiGBgRAyMeJ24Yhz1Gzmw2YzJ3Jnpmsxmb3U7ciCceW6xWLFbLnt/WREIoIiIig4MSOxGRo8QwDCIdEcKhEOFQiPZQe5cEzGqzYXc6sNu9WGw2LBZLIpnbm0ileDy0BXo3xs4wuiZ6RtzAMOLE43ue3+cn8TjW+e9IRwfRaDSx7QAmExaLBZvdjt1hx+ZwYHfYsVitSvhERESSQImdiAxb0UiUWKz71q/+dKjWsWg0SnuwjfY9yVw81vlaq81KiseN3eHA5rBjs9sxm/t32LPJZMJksRzRYOp9E0jDMIjHYkT3dA/t7CYaJRqNEgl30N7WltjPbDZjc9jJ2vPYMAwleiIictSUlpZy5513smrVKm6++WaWLl2K3W4/omP5/X7mz59PJBIhOzubJUuWJJY1CIVCfPOb3+Tee++loKCgy34NDQ3Mnz+f9vZ2cnJyuP/++3G5XKxevZpf//rXWK1Wrr/+eqZPn97n+u5LiZ2IDFuxWPSorC+37wySna1yHYSCbYSCQSLhDgAsFgvOlBScLhcOlxOrzTbgcfUXk8mU6ILZnfieFr5IuIOOcJiO9nBiW9WuMlzulM66p7iU5ImIyFHzyCOP9Gn/J598kq997WvMnj2bZcuW8eKLL3LVVVfx0Ucf8aMf/Yiamppu91u+fDlf/epXmTNnDk8++SQvvvgis2bNYtWqVbz88suEw2Hmzp3LWWeddcRJZ3eU2ImI9JFhGLS3hWgLBGlvCxKLxgCwOx34MtJxut3Y7LZhm9SYzWYcTme3i4zbHHaCLa0Emls6u5J6PXhSU7E7HUmIVEREBrM1a9awdu1a2tvb8fv9XHnllfzlL39h27ZtLFiwgJkzZ/L666/z7LPPYjabmTp1KvPnz6e2tpb58+djGAbZ2dmJ482YMYPXX3+dXbt2sWTJEuLxOC0tLSxevJgpU6Zw7rnnMmXKFHbs2EFmZibLli3DYrEk9r/11ls7e63E41RVVTFq1CgAOjo6eOyxx1iwYEG39Vi/fj3XXnstAP/2b//Gww8/zIgRI5g8eTJ2ux273U5xcTFbtmzhlFNO6bfzp8ROROQI7E3mQsEg7cFdxGIxTCYTzhQXrgw3zpQULFZLzwca5rLz84jH44RD7YSCQdpaAwRbWrE7HHh8qbg87mSHKCIi3Zk1C157rX+Pef758Oqrh3xJMBjkmWee4dVXX+XZZ59l9erVrFu3jpUrV3LaaaexbNkyXn75ZVwuF7fccgvvvvsu77zzDl/96le59NJLee211/jVr37V5ZifffYZCxcuZPz48fzhD39gzZo1TJkyhbKyMp577jny8/O5/PLL+eijj5g0aVJiP5PJRDQa5aKLLiIcDnPDDTcAMHXq1EPWIRAI4PV6AXC73bS2tnZ5bu/zgV6Ome8tJXYiIr3Umcy1dbbMBduIx+OYTCY8aT7sDgfOFFe/j5MbDsxmMy53Ci53CmmZmQRbO1vwGmr9mOvqCQXbyC8uwu5QK56IyLFuwoQJAHi9XsaOHYvJZMLn8xEOh9m9ezcNDQ1cc801QGcSWFZWxrZt27jooosAmDJlygGJXU5ODsuXL8fpdBIMBvHsGUKRnp5Ofn4+APn5+YTDYfZns9l47bXX+J//+R8WLlzI888/32MdPB4PwWAwUV5qamriub2CwWCXRK8/KLETETmEeDze2TIXCBJqC2LEDUx7EpUUjxuny4U7NbXXM1Ue68wWM940Hx5fKuFQ+54Er5bGWj9Z+XnkFhVisw+d8YciIsNWDy1rA+VQwxaKiorIz8/nmWeewWazsWbNGiZMmMD27dv55z//yQknnMBHH310wH733nsvDz30EGPHjuXRRx+loqKix7IA7rzzTr7yla/whS98Abfb3eshFVOmTOGtt95izpw5vP3220ydOpVTTjmFn/70p4TDYTo6OigtLWXcuHG9Ol5vKbETEdlPYsxcayuhYBuGYWA2m0nxeHC53ZoEpB/s7bbqTOmcTKahxo+/sor66hqyC/LJKSro+SAiInJMycjI4KqrrqKkpIRYLEZhYSHnnXce//Ef/8HNN9/Ma6+9RlFR0QH7XXjhhXz3u98lMzOTvLw8Ghsbe1VeSUkJd955J4899hhms5k777zzoK9tampi8eLF/PznP+f6669n4cKFrF69mvT0dH7yk5+QkpJCSUkJc+fOxTAMbr75Zhz93FPFZBiG0a9H7EebN29ONMdK/9P5leFm/2s63N7e61kxDcMgEu4g2NpKWyBIPBbr7ELocZPiceNwHTyZO5y15frqaJXV13Ky8vMAqKuq7vG1ntRUTGYT4fZ2/BVVtDQ2YbZYmHTm6QC0h0JHHMe+LBYrVtvQ+numPqdlONJ1PXjpvTkyoVAosQxCX3X3HvT2fRla33AiIv0sFosRbGkl2NpKtCMCgMvtxp3qwZmSopa5oyAej9PW0plE+jIzcHncNDf866+ptRWVuNx9n2TFm+YbcomdiIhIb+kbTkSOSR3hMIHmFtpaAxiGgd3pID07ixSPG7NFs1kmk93hIHtPix9AXVVN58QrWVlKzERERA5C35AicswwDINQMEigqYVwe/u/1lXz+bA7+m+BUOlfvswMWhoaqd5dRmpGOt40n1pSRURE9qPETkSGPcMwCLa00tLYRCwaxWK14svMwJPqVevcEJCankaKx01TXT3N9Q0EW1vJyMnudkF0ERHpG8Mw9MezJOnr1CdK7ERk2DLiBoGWFloaOhM6u8NBelYmTrfGzg01VpuNrPw8QsEgjf46assrSc1IJzU9Te+liEg/cTqd1NfXk5mZqc/Wo8wwDOrr63H24Y+WSuxEZNgxDIOGWj9Vu8uIhDuwOTrHz2mZgqHP5XbjcDpp9NfT0tBIe7CNjNwcrX0nItIPioqKKC8vx+/3JzuUISUSiWCz9f17yOl0drtcQ28psRORYaW5oZGKHTsJh9pxprjw5edqdsthxmyxkJmXg6s1hUZ/HTVl5aRlZeJO9ep9FhHpA5vNxujRo5MdxpAzWJaJUGInIsNCe1uIjpYA2+u34HA5GT1hPM4UF4HmlmSHJgMkxevBvmdx80Z/HaFgGxm52VjAlKWfAAAgAElEQVQ0blJERI5BSuxE5KiLRqLEYtF+OVYsGsNfVU1DbS0mTOQWFZCRk43JbMaIx/ulDBm8rFYr2QV5BJpbaKpvoGZ3OZl5uThcmlhFRESOLUrsROSoi8WitDY19+kYhmHQ1hqgqb6BeCyGO9VLezSC3ekk0NIKQIrH0x/hyiBnMpnwpvlwuJzUVdVQW1FJWlYmHl+qumaKiMgxw5zsAEREDlc0EsFfWUVDrR+rzUpuUWFnK51u4o9pdoeDvBGFON0pNNXVU19TS1yttiIicoxQi52IDBmGYRBobqa5vhGA9OwsTZghXZgtFrLycmltaqa5voGacAeZeblagF5ERIY9JXYiMiREwh001PrpCIdxpqSQnpOF1aqPMDmQyWQiNT0Nu9NBfXUtteUVpOdk403zJTs0ERGRAaO7IhEZ1AzDoKWxiZaGRsxmMxm5OaR43Gqlkx45XS7yRhRSX11LQ00tJmDEcWN07YiIyLCkxE5EBq1wezuNtX4iHRFSPB7SsjM1lb0cFovVSnZhfmLMXUc4zKjxx2Pth4VkRUREBhNNniIig048Hqeprp7a8krisThZ+blk5uUoqZMjYjKZSM/OIn9UMYHmFj7d8BGhYDDZYYmIiPQrJXYiMqi0t4WoKSuntakZd6qXvOIRuNzuZIclw0B6VibHn3IShhFn64cf01hXn+yQRERE+o0SOxEZFOKxOA21fvyVVQBkF+STkZON2aKPKek/bq+X8ZNOweV2s3PLVip37sIwjGSHJSIi0mcaYyciSRcKBmmsrSMWi+FN85GakY7ZrIROBobNbue4iSdSvn0nNeWVtAXbOsfdaZZVEREZwnq8c4rH49xxxx1cdtlllJSUsGvXri7bV69ezZw5c7j00ktZu3YtAJWVlVx11VWUlJQwb948tm/fDsBf//pXLr74Yi677DJWr149ANURkaEkFotRX11DXVUNZouFnKIC0rIyldTJgDObzRQfN4YRY8cQaGpm64aPCAXbkh2WiIjIEevxz5NvvvkmHR0dvPjii2zYsIElS5awYsUKAPx+P6tWreLll18mHA4zd+5czjrrLH72s58xb948Zs6cyd///ncefvhhHnnkEe6//35eeuklXC4XV1xxBdOnTyc7O3vAKykig4thGARbAzT564jH46RmpJOanqZp6OWoy8rPxel2sWPzVrZu/IiRxx9HWlZmssMSERE5bD3+WXz9+vWcc845AEyaNImPP/44sW3jxo1MnjwZu92O1+uluLiYLVu2sHDhQqZNmwZ0/kXe4XBQWlpKcXExPp8Pu93O1KlTef/99weoWiIyWHWEw5R9tp2GmlqsNht5I4rwZaQrqZOk8aSmMn7SRJwpKezYspXKXbs17k5ERIacHlvsAoEAHo8n8dhisRCNRrFarQQCAbxeb2Kb2+0mEAiQkZEBwPbt21m6dCmPPfYYDQ0N3b72UMLhMJs3bz7sSknvtLe36/zKUWMYBrFwB9G2EGDCbLcRs5ioa+i/mQmjkQhVVVWJxwWFhV0eD5SjVc7RLKuv5WTt+d2bY/RUVn/VtyMaobm19aDbDZsFi8NOTVkF/uoabJ4UTEnuFqzPaRmOdF3LcDNYrukeEzuPx0Nwn/V+4vF4YoD5/tuCwWAieXvvvfe46667eOCBBxgzZgwdHR0Hfe3BOBwOJkyYcHg1kl7bvHmzzq8cFe2hEGXbthMOhvD4fOQVFxIOtfd7OVVVVeTn5yceu1JSujweKEernKNZVn+V05tj9FRWf9XXm+ajoKjokK8xDIP6mlrKt+8gFgwxatzxeNN8/VL+kdDntAxHuq5luBnoa7q3SWOPf4qcMmUKb7/9NgAbNmxg3LhxiW2nnHIK69evJxwO09raSmlpKePGjeO9997j3nvv5amnnmLixIkAjB07ll27dtHU1ERHRwfvv/8+kydPPpK6icgQYRgGNeUVbPnnh4SCQYqPG8txJ0/A7nAkOzSRbplMJrLychl/6kQsFgufffwJ1bvL1TVTREQGvR5b7L70pS/x7rvvcvnll2MYBvfddx+/+MUvKC4u5otf/CIlJSXMnTsXwzC4+eabcTgc3HfffUQiERYtWgTA6NGjufvuu1m0aBHf+ta3MAyDiy++mNzc3AGvoIj0TjQSJRaL9tvx2tvaqNy5m/a2EN40H3nFI7DZbXSEwxjxeL+VIzIQXG434089hbLS7VTtLiPQ0sLIccdhs9uTHZqIiEi3ekzszGYzd999d5fnxo4dm/j3pZdeyqWXXtpl++9///tujzVjxgxmzJhxJHGKyACLxaK0NjX3+TiGYdDS0EhLYxNmi4XMvBxcbjftbW2075lNPmWfcbsig5XFamHkuOPw+lIp276DLf/cSPHxY/FlpCc7NBERkQNoNVYR6TfhUDsNtX6ikQgpXg9pWZlYLJZkhyVyxEwmE5l5uaR4Pez8dBvbP9lCRm42haNHaUFzEREZVPStJCJ9Fo/Haa5vINDcgsVqJSs/D5c7JdlhifQbl9vN+EmnUL27nJryClobmyk+fgyp6Wq9ExGRwUGJnYj0SXtbiIZaP7FoFI8vFV9mBuYkTxEvMhDMZjMFo4rxZWawe9tnlG7qbL0rGj0Ki1rvREQkyfRNJCJHZN9WOqvNSk5hPg6XK9lhiQw4t9ezp/WujJrySlobmykaMwpfZgYmkynZ4YmIyDFKiZ2IHLZ9x9KplU6ORZ2tdyP3tN5tZ8eWrXjTfBSOGYUrRd2QRUTk6NOdmIj0mmEYNNXVU1tRiWEYZBfkk56dpaROjllur5cTJp9C0ZhRtLUG2PLBh5Rv30k02n9Lh4iIiPSGWuxEpFc6wh001NQQ6YjgTvWSlpWphE6Ezpkz9/6Ro3JXGf7KKhpq/eQUFpCW1bfumT6vl3B7+0G3WyxWrDZ9lYuIiBI7EemBYRgEW1ppqqvHZDZrxkuRg7DabBQfN4asvBx2f7adql278VdV4cvIwOVOOaIEr95fh91qO+h2b5pPiZ2IiABK7ETkEGKxGI21fkLBNpwpLjJysjX7n0gPUjweRo0/ntrySpobGqmvrsHucODLzMCZogmGRERkYOgOTUS61R4K0VBdSywWIy0zA0+aTzP+ifSSyWQixevB5XETbG2lpaERf2UVDpcLX2Y6Dqcz2SGKiMgwo8RORLowDIOWhkZaGpuw2mzk5udhdzqSHZbIkGQymfCkpuL2eAi0dCZ4teWVuNwp+DIysDnsyQ5RRESGCSV2IpIQj8Wor6mlvS1EitejGS9F+onJbMab5sOd6qW1qZnWpiZCZeWkeD34MtKx2g4+jk5ERKQ3lNiJCADtbSGqyyqIRaOkZ2fhTvWq66VIPzObzfgy0vH4UmltbCLQ3EJbawCPL5XU9DSNYRURkSOmbxARodFfx65tpZhMJnIKC3C4NP5HZCBZLBbSsjLxpPloaWgk0NxCsKUVj89HaroPs8WS7BBFRGSIUWIncgwzDIPKnbuorajC5XGTnpWpFgORo8hqtZKRk403LY2WhkZam5oItLSQmubDk+ZLdngiIjKE6A5O5BgVi0bZsWUrrU3NZOXnkp2fR6ClNdlhiRyTbHYbmXk5eMM+mhsaaW5opLW5BSxmDMNQt2gREemRZkUQOQZ1hMNs3fgxrc0tFB8/lhFjx2DSJCkiSWd3OMjOzyOnsACb3Ua8I0LVrjKCrQEMw0h2eCIiMojpTk7kGBMKBtn64cd0hDs47qQJZObmJDskEdmPw+UkuyAfs9OO2WKmoaaW2opKwu3tyQ5NREQGKSV2IseQ1qZmtm7cBMC4U07CqzE8IoOWyWTCbLGQW1RIek420UiU2vJK6qtriUaiyQ5PREQGGY2xEzlGNNT62b2tFIfLydiTJmB3aNFxkaGgc5FzLykeN62NTbQ0NRMKBjvXxfN6kx2eiIgMEkrsRAa5aCRKLNa3v87XVdVQW1FJitfDiLGjMQzjgC5dRjzepzJEZGCZzWZ8mRm4U7001TfQ0thEKNjGiLGj8WVmJDs8ERFJMiV2IoNcLBaltan5iPY1DIOWhkZaGptI8bjJyMmmLRDs9rUpHk9fwhSRo8Rqs5GVl0t7KERLfSPbN39Kano6RWNH4XBqDUoRkWOVEjuRYcowjM4p0xubcHu9pOdkacp0kWHE6XKROSGHBr8ff2U1m9dvICs/j8y8HMwDMMutxWLFatNtg4jIYKVPaJFhyDAMmusbaG1qxp3qJT1bSZ3IcGQYBg6nk7ziIprq6vFXVtHoryM9Owtniqtfy/Km+ZTYiYgMYpoVU2SYMQyDprq9SV2qkjqRY4DVaiUrL5fsgjzAwF9ZRUONn3gsluzQRETkKFFiJzKMdCZ19QSam/H4UknPzlRSJ3IMcaakkDuiCG+aj2BrK1W7y2kLaHFzEZFjgRI7kWHiX0ldC540H2lZSupEjkVms5m0rExyRxRisVqor66lrrqGaFRr34mIDGdK7ESGiZaGRgLNLXjTfKRlZiipEznG2R0OcosK8WVmEG4LUb2rjNbmZrXeiYgMU0rsRIaB1qZmWhqbcKd68SmpE5E9TCYTqelp5BUXYXc6afLXU1tRSaSjI9mhiYhIP1NiJzLEBVtbaaqrx+VO0UQpItItq81GdkEeGTnZRDsiVO8up7mhUa13IiLDiOYtFhnCQsE2Gmr8OFxOMnNzlNSJyEGZTCbcqV6cKSk01dXR0tBIWyBARnY2DpcWNhcRGerUYicyRIVD7dRX12Bz2MnKz8M0AAsSi8jwY7FayMzLJSs/DyMep7aikkZ/HfF4PNmhiYhIH6jFTmQI6giH8VdVY7Fayc7Px6ykTkQOk8udgsM1gub6BgLNLYSCbWTkZPf7wuYiInJ06G5QZIiJRqPUVVZjNpnILsjDYrUkOyQRGaLMZjPp2VlkF+ZjMoG/skqtdyIiQ5QSO5EhJB6PU1dVTTweJ6sgD6vNluyQRGQYcLpc5I4owuPzEWhuoXp3Oe1tbckOS0REDoMSO5EhwjAMGmpqiYQ7yMzLxe5wJDskERlGOlvvMskpLMBkMuGvrKah1k88ptY7EZGhQImdyBDR3NBIKNhGWlYGLndKssMRkWHK4XKSO6IQb5qPYEsr1WVlhIJqvRMRGeyU2IkMAcGWVlr3LEDu8fmSHY6IDHNms5m0rExyigowmc3UVVVTsWMX0Wg02aGJiMhBKLETGeTaWgM01HauVacFyEXkaHI4neSNKMKbnkZzfQObP9hAc31DssMSEZFuaLkDkUEs3N5OWekOrDYbmXm5SupE5KgzmUykZWaQmZtD1a7dbN/8KWmZGeQWF2Gx9P+svBaLFatNtyciIodLn5wig1QsFmP7J1swDKNzWYMBuIESEekth9NJdkE+zQ2NNNU30NrcQkZuNk5X/657503zKbETETkC6oopMggZhsHubaW0t4UoGjMKm92e7JBERBKtd50zZ4K/ooqmunoMw0h2aCIixzwldiKDkL+qmqa6evJHjsDjS012OCIiXXTOnFmEO9VLa1MzNWUVdIQ7kh2WiMgxTYmdyCATaGmhYscufBnp5BYVJjscEZFumc1mMnKyycrPJRaLUVNWTktjk1rvRESSRImdyCAS6ehgx5atOBwOiscdp8lSRGTQc7nd5BUX4XKn0FzfgL+iimgkkuywRESOOUrsRAYJIx5n55atxKIxRk8Yh9WqyQNEZGiwWCxk5uWSkZNNRzhMdVk5wZZWtd6JiBxFSuxEBonKXbsJtLRSfNwYXG53ssMRETksJpMJd6qXvOIi7HYHDbV+6qtricViyQ5NROSY0GNiF4/HueOOO7jssssoKSlh165dXbavXr2aOXPmcOmll7J27dou25599lkeeuihxONf/OIXzJo1i5KSEkpKSti+fXs/VUNkaGusq6e2ooqs/M6/eIuIDFVWm43swnx8mRmEgkGqd5fT3taW7LBERIa9Hvt6vfnmm3R0dPDiiy+yYcMGlixZwooVKwDw+/2sWrWKl19+mXA4zNy5cznrrLOIx+MsXryYjRs3cu655yaOtWnTJpYuXcrJJ588cDUSGWLC7e3s3lZKitdD4ehRyQ5HRKTPTCYTqelpOFNc1NfU4q+sxpuWhi8zXWOHRUQGSI8tduvXr+ecc84BYNKkSXz88ceJbRs3bmTy5MnY7Xa8Xi/FxcVs2bKFcDjM7Nmzue6667oca9OmTTz55JNcccUVPPHEE/1cFZGhp3Nc3TZMwOjx4zCb1TtaRIYPu8NBblHhnmURmqgtr9TEKiIiA6THFrtAIIDH40k8tlgsRKNRrFYrgUAAr9eb2OZ2uwkEAvh8Ps4++2zWrFnT5VizZs1i7ty5eDwebrzxRtauXcv06dMPWnY4HGbz5s1HUi/phfb2dp3fJIsEQ8Taw9g8KZTu6L5rss/rpd5fN+CxFBQWUlVVNeDlDGRZ0Uiky3GPVp2Gw7nr73Ky9vzuzTF6Kqu/6tsRjdDc2tovx+pJf/1/u/81vb+hdO2ZHXY6wmGqdpVhdtgxWy3dvu5ovk+SHLr/kOFmsFzTPSZ2Ho+HYDCYeByPxxOz9e2/LRgMdkn09mUYBt/4xjcS26dNm8Ynn3xyyMTO4XAwYcKE3tVEDtvmzZt1fpOotamZzz7+hMzcHIqPH3vQ14Xb27FbbQMejyslhfz8/AEvZyDLqqqq6nLco1Wn4XDuBqqc3hyjp7L6q77eNB8FRUX9cqye9Nf/t/tf0/sbatdeNBKhvrqWjnAYl8NLWlbmAT0Vjub7JMmh+w8Zbgb6mu5t0thjv68pU6bw9ttvA7BhwwbGjRuX2HbKKaewfv16wuEwra2tlJaWdtm+r0AgwFe/+lWCwSCGYbBu3TqNtZNjViQSYefWbThdLorGjEp2OCIiR4XVZiOnqABvWhrBllZqyyuJdKhrpohIf+ixxe5LX/oS7777LpdffjmGYXDffffxi1/8guLiYr74xS9SUlLC3LlzMQyDm2++GYfD0e1xvF4vN998M1deeSV2u50zzjiDadOm9XuFRAY7wzDYvfUzYpEoY0+cgNnSfXckEZHhyGQykZaVgcPlpKGmlprycjJyckjxaJkXEZG+6DGxM5vN3H333V2eGzv2X93GLr30Ui699NJu950zZ06Xx7Nnz2b27NlHEqfIsFFXVU1LYxNFY0bpRkZEjlkudwq5Iwqpr66lvrqGjjQfvsyMZIclIjJkaQo+kaOoLRCkYscuUtPTycrPS3Y4IiJJtbdrpseXSmtTM/6KKnXNFBE5Qj222InIgaKRKLFY9LD2icfj7NyyFYvVQl5xER3hcK/2M+LxIwlRRGRIMJlMpGdnYXc6aaz1s/2TLYyZMB6PLzXZoYmIDClK7ESOQCwWpbWp+bD2aaqrJ9zeTnZBHqF9ZpPtSco+y42IiAxXbq8Hu8NOQ42fbR9/wogxo8nKz012WCIiQ4a6YoocBeFQO61NzbhTvThTUpIdjojIoGSz2xk9YRypaT7KSrdT9tl24uq1ICLSK0rsRAZYPB6nodaPxWolLSsz2eGIiAxqFquVMSeeQE5RAXXVNZR+/InG3YmI9IISO5EB1lzfQDQSISMn+4CFeEVE5EAmk4nCUSMZOf54goEAn364kbZA77uwi4gcizTGTmQAtYdCBJpb8PhScaa4kh2OiMigZ8QNwu3tQOe4u1Hjx1FWup2tGz+icPQoUtPT+q0si8WK1aZbIREZHvRpJjJA4vE4DTWdXTC1NpOISO/E43HaWgJdnsspyKeuqoby0h34MjPwpvkwmUx9Lsub5lNiJyLDhvqFiQyQ5voGYtEombnqgiki0hcWq5XswnxcHjfN9Q00+uswDCPZYYmIDCq62xQZAO1t/+qC6XCpC6aISF+ZzWYyc3PwpqURbGmlrqpaM2aKiOxDiZ1IP9s7C6bVZlMXTBGRfmQymUjLyiA9O4v2thC15ZVEo9FkhyUiMigosRPpZy0NjcSiUdJzstQFU0RkAHh8qWQX5BGNRqgtq6AjHE52SCIiSae7TpF+1BHuoLWpmRSvB6e6YIqIDBhnSgo5hYVggtqKStrbQskOSUQkqZTYifQTwzBo9Psxm81aiFxE5CiwO+zkFBVitVrxV1bR1hroeScRkWFKc/yK9JNgSysd7WHSc7KxWCzJDkdE9rPv+mgDX5Ym9TharFYrOYUF1FXVUF9TSywWxZvWf2vdiYgMFUrsRPpBLBajub4Bh9OJ2+tJdjgi0o3u1kcbKCkefQ4cTWaLheyCPOpr/DTVNRCLxvBlZvTLWnciIkOFumKK9IOmunri8TjpOVm6kRARSQKT2UxmXg4eXyqtTc001Pi11p2IHFPUYifSR+1tIdpaA3jT07DZ7ckOR0TkmNW5HEImFquV5voGYrEYWfm5mqFYRI4J+qQT6YPOCVPqsFitpKZrTIeISLKZTCZS09PIyMkmHOpc6y6mte5E5BigxE6kD1oam4hGIqRna806EZHBxJ3qJSs/j2gkQk15JZGOjmSHJCIyoHQnKnKEopEorY1NuNxuXO6UZIcjIiL7cblTyCkswDDi1JZXHrVZUUVEkkGJncgRaq5vwADSsjKSHYqIiByE3ekgp7AQs8WMv6KKUDCY7JBERAaEEjuRI9AWCNIWCOBN82G12ZIdjoiIHILNbiOnsBCr3UZdVQ2BlpZkhyQi0u+U2IkcJsMwqCmrwGyxaMIUEZEhwmK1kFNYgDPFRWNtHc0NjVoOQUSGFSV2Ioepqa6eUDCILzNdE6aIiAwhZrOZrPw8UrweWhoaqd5dpuRORIYNrWMnchjisRgVO3fhdLlwe73JDkdERA6TyWQiIycbi8VKo7+eeCzOqPHHY7ZYkh2aiEifqLlB5DDUVlYRCXeQO6IQk8mU7HBEROQIdC5knkFecRHNDY189vFmopFIssMSEekTJXYivRTp6KCmvAJfRjruVLXWiYgMdRk52Yw6YRxtgQBbN27ScggiMqQpsRPppapdZRhxg8LRI5MdioiI9JP0rEzGnjyBaKSDTzd8RGtTc7JDEhE5IkrsRHqhLRCkvqaW7II8HC5XssMREZF+5PX5GH/qRGw2G59t2oy/qjrZIYmIHDYldiK9ULFjJ1arldwRRckORUREBoDD5WLcqSeTmuajvHQHZaXbMeLxZIclItJrSuxEetDa1EyguYXc4iKsVk0kKyIyXFmsVsaceAI5hQXUVdXw2SZNqiIiQ4cSO5FDMAyDyp27sTnsZOXlJjscEREZYCaTicLRIyk+/jiCLa18+uFHtAWCyQ5LRKRHSuxEDqGloZG2QID8EUVajFxE5BiSmZvN8RNPwogbbP3wI+qqqrWYuYgMaupXJnIQhmFQuasMh9NJRm5OssMREZF+ZsSNQy5xYLXbGD1hHBU7dlFWuoPmhkbyRxVjOczFzC0WK1abbrlEZGDpU0bkIBrr6mlva2PU+OO1GLmIyDAUj8dpawn0+Lr07CysNhvN9Q20BQJk5uVidzh6XY43zafETkQGnPqWiXTDiMep3lWGMyWFtKzMZIcjIiJJZDKZSE1PI6cwHyNuUFNeSaC5RV0zRWRQUWIn0o36Wj/h9nYKRo5Qa52IiACdSyLkFhfhdDlp9NdRV1VDLBpNdlgiIoASO5EDxONxqsvKSfF6SM1IT3Y4IiIyiFgsFrLy80jLyiQcClG1u5xga6ta70Qk6ZTYieynrqqGSLiDgpHFaq0TEZEDmEwmvGk+ckcUYrPbaKjxU1+t1jsRSS4ldiL7iMVi1JRX4PGl4k3zJTscEREZxGx2OzmFBfgyMwi1hajeXU5ba0CtdyKSFErsRPbhr6wiGolQMLI42aGIiMgQsHdilbwRhVhtNupraqmrqibSEUl2aCJyjFFiJ7JHLBqjtqKS1PQ03KneZIcjIiJDiM1uJ6eoYM/Yu3aqd5fRXN9APB5PdmgicozQoioie9RVVxOLxsgrLkp2KCIiMgTtHXuX4nHTVN9AS2MTwdYAZpMJe16uxm2LyIBSi50IEI/FqK2owpvmw+1Va52IiBw5i9VKZm4O2YX5mMwmykp3sP2TLYTa2pIdmogMY0rsRIC6mlqikQh5I9RaJyIi/cPpcpE3oojcogICLa1s+eBDdm8rpSMcTnZoIjIM9ZjYxeNx7rjjDi677DJKSkrYtWtXl+2rV69mzpw5XHrppaxdu7bLtmeffZaHHnoo8fivf/0rF198MZdddhmrV6/upyqI9E08Hqe2vBJPqhePLzXZ4YiIyDBiMpnIzMvlpNMmk12QT0Otn0/W/5PKnbuIankEEelHPY6xe/PNN+no6ODFF19kw4YNLFmyhBUrVgDg9/tZtWoVL7/8MuFwmLlz53LWWWcRj8dZvHgxGzdu5NxzzwUgEolw//3389JLL+FyubjiiiuYPn062dnZA1tDkR401PqJdHRQfPzYZIciIiLDlNVmo2jMKLIL8qjaVUZNeSV11bXkjSgkKy8Xs8WS7BBFZIjrscVu/fr1nHPOOQBMmjSJjz/+OLFt48aNTJ48Gbvdjtfrpbi4mC1bthAOh5k9ezbXXXdd4rWlpaUUFxfj8/mw///27jxGzvq+4/h77uuZ+9rLa3ttLxCoRUwSpQ2gBpoeaWhVEiHi1jRtEgVElYSGFpo0lAYXSKtEqRKRNGmbSm4STEsr9R+kirbCLWlQMLgE4gMMPnZ3dufYuWfnfvqHYYMT4921dz07s5+XZHlnnmee3/eZ+e3O83mun9PJNddcw7PPPrsGqySyfGa3y9zpabyGoXHrRERkzbncbrZctoPLrt6J1/Ax/dpJXnr2edLTM3Q7nV6XJyJ9bMlgV6lUMAxj8bHNZls8daBSqeB/040mfD4flUqFYDDItdde+zPLOde8Ir00n8nRbDQY2jSqu5WJiMgl4zV8bL/qbez4uStxez2LAW9uaoaOAp6IXIAlT8U0DINqtbr4uNvtYrfbzzmtWq2eFd7Ot5zzzS2dz30AACAASURBVPuGRqPB4cOHlypRLlC9Xt/Q769pmjQLZSw2K9Nzs8yk55b92qDfTy6TXcPqfmJkdJRUKjUw7axlW+1W66zl6r3rXTux1/9fzjKWamu11rcfP6ef7tNr1c5y9EvfW49tWWxWctm3/s7whYM4PG6qhSIzJ06SOnUaX9CPx29gta7sPnddoFwuX2TFa2ujb3/I4FkvfXrJYLdr1y7+67/+i/e///0cOnSIycnJxWk7d+7kK1/5Co1Gg2azyfHjx8+a/mbbtm3j5MmTFAoFvF4vzz77LB/96EfP27bL5eKKK65Y4SrJch0+fHjg3t92q02ns7yL0Yvzeabni4xNbCEQCa+oHbPbxWl3XEiJK+bxehkeHh6YdtayrVQqddZy9d71vp3lLGOptlZrffvxc/rpPr1W7SxHv/W99dSWx+PF7Jx/oHKn4cAwDBoLdUr5PJV8kWqxjBEM4A8GsNmXN/SwPxRkbGx93+F5ELc/ZGNb6z693NC45F+J973vfTz99NPceuutmKbJgw8+yLe//W3Gx8e58cYb2bNnD7t378Y0Te666y5cLtc5l+NwOLj33nv56Ec/immafPCDHySZTK5srUSW0Om0KReKS85nmibpqRnsDgdYLMt6zZt533R6soiIyGpxedzEPcM06w1K+QLlfIFyoYjP7ycQDp753hIROYclg53VauULX/jCWc9t2/aTuwfecsst3HLLLed87c0333zW4xtuuIEbbrjhQuoUWVX1Wo1Ws0kkEde1dSIisu443S5iw0lazSblQpFqqUS1VMJr+PCHQzjfYke6iGxcyzuuLzJATNOkNF/AZrfj9evIm4iIrF8Op5NIIk4wEqZcKFIplqhVqri9HvyhEC6PWzsoRQRQsJMNqFmv02w0CMdj+jIUEZG+YLPbCcWiBMIhKsUS5WKJzEwKp8uFPxzC4/PqO01kg1Owkw2nVChitVp1tE5ERPqO1WYjEAljhILUyhVK+QK52TnsDgf+cAgjcP47jovI4FKwkw2l1WxSr9YIhEMrvoW0iIjIemG1WjGCAXwBPwuVKqV8gXw6QzlfIDk2QjSZxGa39bpMEbmEFOxkQykXimCxYASDvS5FRETkolksFrx+A4/ho15boFapMP3aSWZPTxEbHiI+MoxDd9IU2RAU7GTD6LQ71MoVfH5DezFFRGSgWCwWPD4vidFh2q0Wc1MzzJ2eJj2dIppMkBgdxuV297pMEVlDCnayYVSKRUzTxB/S0ToRERlMZtfE7nAwunUzsaEkudk5srNzZFOzBCNhokNJ3F7PRbdjs9mxO7QZKbKe6DdSNoRut0ulWMLt9eJwOntdjoiIyJrodrvUSpXFx/5wCI/foFwoUsoXKM7ncXs9BMIhXJ4LD3j+UFDBTmSd0W+kbAi1coVut4s/rKN1IiKysdjtdsJvGiqhUiiSnk7hcrsJRM4EPA2VINL/FOxk4JmmSblQxOly6foCERHZsGw2G8FIGH8oSLVUppwvkJmZxelyEYiEcXsV8ET6mYKdDLx6tUa71SKaTOgLS0RENjyr1Yo/FMQIBqiWypTyBbKpWRwuJ4FwWIOdi/QpBTsZeKVCEZvdjsfw9boUERGRdcNisSyOhVctnzmCl5udw+F0nAl4hk8BT6SPKNjJQGvU6zTrdUKxqL6cREREzsFisWAEAvj8fmqVCqX5Arm5NPZ5B4FwCK/f0HeoSB9QsJOBVs4XsVit+AL+XpciIiKyrlksFnx+P17DYKFapTRfYD6doTifJxAO4fMbWKzWXpcpIm9BwU4GVrvVYqFaxR8KYdUXkYiIyLJYLBa8hoHH56Neq1GcL5DPZCnO5/EHA/iCgV6XKCLnoGAnA6tcKALgD+kLSEREZKUsFgsenw+310tjoU65cGYcvFK+QL1aY2h8THebFllHFOxkIHU6HaqlMl6/gc2ubi4iInKhLBYLbq8Ht9dDs9GkXCgwn8kyn84QikWJDw/hC/h1HZ5Ij2mLVwZStVjCNE38oVCvSxERERkYTpeTaDLB6FYPxVye3OwchWwOt9dLfGSIcDyGzWbrdZkiG5KCnQwc0zQpF0u4vR6cLmevyxERERk4DqeT0a2bGRofI5/Jkk3NcvqVV5l57SSRZILYcBK3x9PrMkU2FAU7GTjVcplup4M/FOx1KSIiIgPNZrMRG0oSTSaolstkU7NkUrNkZlIYgQDRoQShaASrjuKJrDkFOxkopmlSLhRxOJ24tKdQRETkknhjLDwjEGB0a5PcXIbc3Bwnj73ClM1GOBEjmkzgNYxelyoysBTsZKBUSiXazRaRRFwXcYuIiKwRs2vSqNffcno4HiUUi1ArVyhkc+Rm02RTc7i9HhxuF7VKZdk3N7PZ7Ngd2mQVWYp+S2Sg5GbT2Gw2vH7tERQREVkr3W6XWqmyrHkDkTBGMEC1XKFaLlOfL3AsX8Tj8+EL+HF53OfdGesPBRXsRJZBvyUyMGqVCrVyhWA0oqN1IiIi64jVZsMfCuIPBZmZmsbjclEtV6hVKtgddnx+P76AX0MUiVwE/fbIwEhPp7BarRgBDUguIiKyXllsVsLxGMFohIVqlWqxTHE+T3E+j9vrxQj4cfu82kkrskIKdjIQmo0G+UyWSDKO1WbtdTkiIiKyBKvVeuZInd9Pq9miWiqfubPmbA2rzYbPb+DTzlqRZVOwk4GQmUkBEEkkaCws9LgaERERWQmH00EoFiEYDVOv1aiUypQLRcqFIqX5PPGRIQ2bILIEBTvpe512m+xsmnA8itPlVLATERHpUxaLBY/Ph8fno9NuUy1XWKhUfzJsQjxGdEjDJoici4Kd9L3sbJpup0NidKTXpYiIiMgqsdntBMIhRraM02o2yc2myaXTZGfn8Ph8xIaTROIxHcUTeZ2CnfQ1s9slM5PCCAbwGsZ5x9QRERGR/mOxWPAHg/iDQdrtreQzWbKpOU6/8iozr50kkkwQG07i9nh6XapITynYSV/LZ3O0mk02bZ/odSkiIiKyxux2O/HhIWJDSaqlMpnU7Jl/Myn8oSDx4SECkbDuqCkbkoKd9C3TNElPz+D2eAiEQ70uR0RERNaA2TXPeUaOw+VkZMs4idFh8pkc+UyWVw8fxeF0Eo5HCcWi2B2OZbdjs9k1ELr0NfVe6VuVYomFao3x7RPaMyciIjKgut0utVLlvPO4vR6GxsdYqFapFEukp1Okp1N4/QZGMIDT5VpyW8EfCirYSV9T75W+NTc9g93hIJyI97oUERER6TGLxYLXMPAaBq1mk0qxRLVUplau4HA5F6/Ht1o13q0MJgU76UsL1RrlfIHh8U36Ay0iIiJnOXM6ZoxgNEK1XKZSLJFPZylm5/EF/BjBwIpO0xTpBwp20pfS0zNYrFZiw8lelyIiIiLrlNVqxR8MYgQCNOp1KsXS4sDnbq8HIxjE7fXokg4ZCAp20ndazSb5TJboUEJ720RERGRJFosFt8eD2+Oh025TKZaolMpkU7PY7HaMYACP4cPV60JFLoKCnfSdzMwspmmSGNGA5CIiIrIyNrudYDRCIBJmoXLmZivF3Dyl+TzheIzY8BA+v9HrMkVWTMFO+kqn0yGbmiUUjeDyuHtdjoiIiPQpi8WC12/g9Rs0G02a9TqFbI75dAav4SM2PEQ4HtO1/NI3FOykr8zPpel0OiRGdbROREREVofT5SSajLNp21bm0xkyqTlOvXyc6ddOEk0miA0ncbm1Q1nWNwU76Rtmt0t6OoXP78cX8Pe6HBERERkwNrud+MgwseEhKsUSmdQs6ekZ0tMzBMJhYsNJAuGQbrYi65KCnfSNfDZHs9FgbGJLr0sRERGRAWaxWPCHgvhDQZqNBtnZOXKzaV79cR6ny0V0KEk0mcDh1E3cZP1QsJO+YJomc1PTuL0eApFwr8sRERGRAWN2TRr1+jmnRZMJIvEYpUKRfCZL6uQpUqdOEwgHCcdjeA1j2UfxbDY7doc2wWX1qVdJXyjO56nXFtg8uV2nP4iIiMiq63a71EqV885jtVqJJhMEwqHXx8QrUZovYHc4MIIBfH4Dq8123mX4Q0EFO1kT6lWy7r1xtM7pchGOx3pdjoiIiGxwDqeTcDxGMBqhVqlSLZYoZHMUc/N4DQNf0K+brcglp2An616lWKJWrjC2bauO1omIiMi6YbVaMQJ+jICfZr1BpXRmm6VaLuNwOTGCAbyGoSET5JJQsJN1b25qGrvDQTSZ6HUpIiIiIufkdLuIuOOEolGqlTLVYol8OksxO48vGMAIBrDbtekta2fJ3tXtdrn//vs5evQoTqeTvXv3snnz5sXpjz32GI8++ih2u5077riD9773vczPz3P33XdTr9dJJBI89NBDeDwe9u7dy3PPPYfP5wPgkUcewe/XbevlrdXKFcqFIiNbxrW3S0RERNY9q82KPxjECARo1OtUCkXK+QLlfAGv4cNmt+s0TVkTSwa7J598kmazyf79+zl06BAPP/wwX//61wHIZDLs27ePxx9/nEajwe7du3nPe97DI488wgc+8AFuvvlmvvnNb7J//34+8pGP8NJLL/G3f/u3RCKRNV8xGQyzU9PYbDZiQ8lelyIiIiKybBaLBbfHg9vjod1qUS6WqJZKnDhyjLRhEB8ZIhSLase1rJole9LBgwe57rrrALj66qt58cUXF6e98MILvP3tb8fpdOL3+xkfH+fIkSNnveb666/n+9//Pt1ul5MnT3Lfffdx66238s///M9rtEoyKOq1BYq5eWIjQ9h06oKIiIj0KbvDQTgWZWTLZobGx+i025w89go/fvY5Zk9P0Wq1el2iDIAlt5YrlQqGYSw+ttlstNtt7HY7lUrlrFMpfT4flUrlrOd9Ph/lcplarcbv/M7v8Hu/93t0Oh1uu+02rrrqKi6//PK3bLvRaHD48OGLWT85j3q9vq7f31alBkChUqa4zDqDfj+5THYtywJgZHSUVCq15u1cyrYGYZ3ardZZy9V717t23rh/7XKWsVRbq7W+/fg5/XSfXqt2lqNf+t56bEvrdLal+vVqtrUSl6qdaDwGXhcOh41OvUHq5GlSJ09jczmxuV1Y7ecfLkHWn/WyTb1ksDMMg2q1uvi42+0uXvj509Oq1Sp+v3/xebfbTbVaJRAI4PF4uO222/B4PAC8+93v5siRI+cNdi6XiyuuuOKCV07O7/Dhw+v2/W3WG7x08Hniw0OMbdu67Nc16nWcdscaVnaGx+tleHh4zdu5lG0NwjqlUqmzlqv3rvftLGcZS7W1Wuvbj5/TT/fptWpnOfqt762ntrROZ1uqX69mWytxqdrxh4KMjI0tPl6o1cjMzDKfztBpNPGHgiRGR/CHgrobeJ9Y623q5YbGJU/F3LVrFwcOHADg0KFDTE5OLk7buXMnBw8epNFoUC6XOX78OJOTk+zatYunnnoKgAMHDnDNNddw4sQJdu/eTafTodVq8dxzz3HllVdeyLrJBpCengEgMTrS40pERERE1o7H62V8+wRXvXMXw5vHWajVOP7SYY48/wK5uTTdbrfXJUqfWPKI3fve9z6efvppbr31VkzT5MEHH+Tb3/424+Pj3HjjjezZs4fdu3djmiZ33XUXLpeLO+64g3vuuYfHHnuMcDjMl770JbxeLzfddBO33HILDoeD3/zN32THjh2XYh2lzzQbDbKzc0QSMZxuV6/LEREREVlzdoeDoU2jJEaHyWeypKdTnHr5ODMnTxEfHiY2lMDuWPuzkqR/LRnsrFYrX/jCF856btu2bYs/33LLLdxyyy1nTY/FYvzd3/3dzyzr4x//OB//+McvtFbZIGZPTQEwvGlTjysRERERubSsVivRZIJIIk65UCQ9PUPq5CnmTk8RSSZIjAzj8mi4BPlZutWgrCv12gK5uTTxkSEdrRMREZGBY3ZNGvX6suZ1edxs2j6xuH2UnZ0jm5rFHwoSHUriNXznfb3NZsfu0Ob+RqFPWtaV1KnTWK1Wkm+6qFhERERkUHS7XWqlyopfFwiH8PkNysUSlWKJcqGI0+3CHwri8fnOeaMVfyioYLeB6JOWdaNWqVDI5hjaNIrDqXPIRURERN7MZrcTikYIhENUS2XKhSK52TQ2ux1/KIgv4NeA5xuYgp2sGzMnTmGz23UnTBEREZHzsFqt+ENBjGCAhWqVcr5IIZujNJ/HFwzgDwaw2bWZv9HoE5d1oVwsUi4UGdmyWX+IRERERJbBYrHgNQy8hkFjoU65UKCcL1AuFDECfjw+Ly63brSyUWgLWnrONE1SJ07hcDqJDyd7XY6IiIhI33F53Lg8Q7SaLUr5PJViiZd/9GPiw0Mkx0Y0VMIGoGAnPVeaz1MtV9i0fQKrzdbrckRERET6lsPpIJpMEAiHWKhUSU/PkJ2dJT4yTGJ0BLvOjBpY+mSlp0zTZObkKVxuN9FkotfliIiIiAwEh9NJZCLO8JZxZk9NMXd6mmxqjtEt40SSiXPeRVP6m26bIz2Vz2Sp1xYY3rxJf2BEREREVpnH62Xr5ZNcdvVO3F4Pp155lZdfeImFarXXpckq0xE76ZlOu8PMyVN4fD5CsWivyxEREREZKG8eDN1mtzG+YxvF3DxzUzMcef4FIsk48ZFhbBd5KYwGQl8f9AlIz6ROnaLVaLL1skkdrRMRERFZZecaDN1mt5PcNEoxN8/8XIZidp5QPIbX8F1wOxoIfX3QJyBrrt1q0+m0z3puoVolMzNLOB7D7nQs7k26WGa3uyrLERERERlUNpuNSCKOL+Ann86Sm52jEQwQikawaIDzvqVgJ2uu02lTLhQXH5umydzpaWw2G16/cda0i+U1jFVbloiIiMggc7ndi0fvyoUijXqd2FBSQyP0KUVyueTKhSKtZpNQPIZVe4VEREREesZisRCKRYkNJ+m02syenqJWqSz9Qll3tFUtl1S71aI0n8fj817UudwiIiIisno8Ph/JTaM4HE5ys2nymSymafa6LFkBBTu5ZEzTJJ/OAhCKx3pcjYiIiIi8md3hIDE2ghEMUCmWSE/N0Ol0el2WLJOCnVwytUqV+sICwWgEu12Xd4qIiIisNxaLhXA8RnQoQavZJD01Q7vV6nVZsgwKdnJJdDodCtksTpcLIxjodTkiIiIich5ewyA+Mky302FuaoZmo9HrkmQJCnay5kzTpJDJ0e10CSdiGrNOREREpA+4PG4SoyNYLJCemqFeW+h1SXIeCnay5ubnMtQqFQKRME6Xq9fliIiIiMgyOVxOEmOj2Bx2MjMpamXdMXO9UrCTNVXKF5ibmsbj8xIIh3pdjoiIiIiskN1uJzk6gtPtJjeXXtUxiGX1KNjJmqnXFjhx5Bguj5tIMqFTMEVERET6lNVmIz4yhMfnpZDNKdytQwp2siba7TavHj6CxWph0/ZtGohcREREpM9ZrVaiQ0k8Pp/C3TqkrW1ZdaZpcuLIMRr1Blsvvwyny9nrkkRERERkFVgsFqJDCR25W4cU7GTVTb92knKhyKZtWzW0gYiIiMiAORPukovhbn4u0+uSBAU7WWXZ2TkyMyniw0PEhpK9LkdERERE1sCbw93s6SkyM6lel7Th2XtdgAwG0zSZm5omdfI0/lCQ0a2be12SiIiIiKyhN8JdMTfP1KsnAAvxkaFel7Vh6YidXDSz2+X0K6+SOnmacDzGxNsux6KbpYiIiIgMPIvFwtjEFoKRMFOvvkYmNdvrkjYsbX3LRWm327zy0hFyc2mGNo2yeXK77oApIiIisoFYrFa2XD55JtwdV7jrFW2BywVr1hu8/MKLVEolxndsY3jzuMaqExEREdmArK+Hu8Dr4S6rcHfJKdjJBamWyhz9vx/RajTZfuUVRJOJXpckIiIiIj1ktVrZevkkgXCY08dfI5ua63VJG4puniIr0mq1SJ04RW4ujdPlYuLn3obH6+11WSIiIiKyDlitVrZeMclrh49y+virYEF3Sr9EFOxkWUzTJJOaZfbkaTqdDvGRYYbGx7Db1YVERERE5CfOhLvLzoS7V14FE2LDCndrTVvlsqRyocjUq69Rry2cGcpgYouO0omIiIjIWzor3B1/lXa7RXJsVPdjWEMKdnJOpmlSLhTJzKQo5Qs4XS62XnEZwUhYv5AiIiIisqQ3wt2pl4+TOnmadrPF6MQWbUuuEQW7Dczv99Oo1896rtPpUMjmyKezNBsNbHY78dFhoskEVquVZqOx4nbMbne1ShYRERGRPmK1Wtk8uR27w0FmJkWr1dLwWGtEwW4Ds3LmNEuAZqNJtVSiWipjmiZOl4tIMo7XMLBYLFRL5Qtux2sYq1SxiIiIiPQbi8XC6NbNOJwOZk6cot1qM3HFZdjstl6XNlAU7DawTqdDuVKjWi7TajQB8PoN/MEgTrerx9WJiIiIyKCwWCwkx0axO5ycevkVXv7RS2y78nIcTmevSxsYCnYbTLfToTifZz6doZQvAOBwuQjFonj9Bjab9pyIiIiIyNqIJuPYHXZOHDnG0f/7EVsu24ERCPS6rIGgYLcBmKZJtVRmPp0hn83R7XRwOJ14g34CwaD2lIiIiIjIJROMhNnxc1fy2pFjvPzCSwyNjzG0aUw3VblICnYDrL6wQD6dYf71G6FYrVZCsSiRRBwjGCA1PY3D7uh1mSIiIiKywXj9Bpe/fSenj7/G7KkpyoUiWyZ36HKgi6BgN2DarRb5bI58OkO1XAHAHwoyvHkTwWhEp1qKiIiIyLpgs9vZctkOAuEQp4+/ypHn/49NO7YRjkV7XVpfUrAbAN1ul1K+cOa6ufk8pmni9noY2TJOOB7D6dKeDxERERFZnyKJOD6/nxNHX+bEkWMU4zGGN2/C5Xb3urS+omDXp0zTpFIsnRlzLpuj025jdziIDQ8RScTx+Lw6T1lERERE+oLL42Zy55WkTk+Rnpohn80RTcYZGhvT6ZnLpGDXR0zTpFwsUsjmKObmabfaWK1WApEwkUScQDikMCciIiIifclitTKyeZzYUJK5qRlys3PMz2WIDiVIjo3qLLQlKNitc81Gg3KhSLlQpJQv0Gn/JMyFY1EC4RBWXTcnIiIiIgPC6XKxadtWkqMjzE5Nk51Nk5tNE0nECcUiGMEgVqu112WuOwp264hpmrRbLaql8pkwVyzSWKgDYLfbCYRDhGJRAqGgwpyIiIiIDDSn28X49gmSYyPMnZ4mn82Sm0tjtdkIRkIEIxEC4TA2u7aLYRnBrtvtcv/993P06FGcTid79+5l8+bNi9Mfe+wxHn30Uex2O3fccQfvfe97mZ+f5+6776Zer5NIJHjooYfweDznnHcjMk2TTrtNo15noVqjXq2xUKuxUK3RabcBsNqsGIEAsaEk/lAQt1fXzImIiIjIxuNyuxnfsY2xbVspF4oUc/MU5+fJZ3JYLBZ8AT8erxe3z3vmf68Hm33jHb9aco2ffPJJms0m+/fv59ChQzz88MN8/etfByCTybBv3z4ef/xxGo0Gu3fv5j3veQ+PPPIIH/jAB7j55pv55je/yf79+/n1X//1c87r7LPBsd8IZaZpYnbNM/+bJqbZpds16XY6dDodOu3Om35u0Ww0aTV/8s/smovLtFqtuH1eQtEIHp8Xj2HgM3xYdIhZRERERAQ4s80cjIQJRsKY5gTVUplCbp5quUxuLk23212c1+Fy4nK7sTscr/+zL/5ss9uwWqxYrVYsVitWqwWL1YrD6ezrAylLBruDBw9y3XXXAXD11Vfz4osvLk574YUXePvb347T6cTpdDI+Ps6RI0c4ePAgn/jEJwC4/vrr+fKXv8ymTZvOOe/OnTvXaNXWRno6xcyJkyt6jcVqweF04nA68RoGTteZn50uFx6fD6fb1dedSERERETkUrJYLBjBAEYwAJw5+NJsNF4/E26Beq1Ks95koVKl3WrR6XSWXGY0mWB8x7a1Ln3NLBnsKpUKhmEsPrbZbLTbbex2O5VKBb/fvzjN5/NRqVTOet7n81Eul99y3vNpNBocPnx4xSu11tzR0AW9rgN0zA71+gLUF848mVm9utat/PzgtaV1Wpdt2d0uMm9ett67nrWT+fGPl7+ct5jnx68vI7Na69yHn9PP9Ok1amdZ+qTvrcu2tE5nWbJfr2JbKzJo7VzCtlbt7/Rqc9qwOG04cONYxuzVdvOCs8daZpZGo7Gs+ZYMdoZhUK1WFx93u13sr5+z+tPTqtUqfr9/8Xm32021WiUQCLzlvOdz9dVXL2slRERERERENrIlL+LatWsXBw4cAODQoUNMTk4uTtu5cycHDx6k0WhQLpc5fvw4k5OT7Nq1i6eeegqAAwcOcM0117zlvCIiIiIiInJxLKZpmueb4Y27Yh47dgzTNHnwwQc5cOAA4+Pj3HjjjTz22GPs378f0zT5xCc+wa/8yq+QzWa55557qFarhMNhvvSlL+H1es85r4iIiIiIiFycJYOdiIiIiIiIrG+6n76IiIiIiEifU7ATERERERHpcxtvSPYN7o1rJo8ePYrT6WTv3r1s3ry512WJrFir1eKzn/0s09PTNJtN7rjjDrZv3869996LxWJhx44d/Nmf/RlWq/ZfSX/J5XLcfPPN/P3f/z12u119Wvra3/zN3/Cf//mftFotPvzhD/Oud71LfVr6WqvV4t5772V6ehqr1coDDzywbv5W6zdpg3nyySdpNpvs37+fz3zmMzz88MO9Lknkgvzbv/0boVCI7373u3zrW9/igQce4KGHHuLTn/403/3udzFNk//4j//odZkiK9Jqtbjvvvtwu90A6tPS15555hmef/55vve977Fv3z5mZ2fVp6XvPfXUU7TbbR599FHuvPNOvvKVr6ybfq1gt8EcPHiQ6667DjgzTuCLL77Y44pELsyv/uqv8qlPfWrxsc1m46WXXuJd73oXANdffz3f//73e1WeyAX54he/yK233koikQBQn5a+9j//8z9MTk5y5513cvvtt/OLv/iL6tPSm47j2gAABbRJREFU97Zu3Uqn06Hb7VKpVLDb7eumXyvYbTCVSgXDMBYf22w22u12DysSuTA+nw/DMKhUKnzyk5/k05/+NKZpYrFYFqeXy+UeVymyfP/yL/9CJBJZ3PkGqE9LX8vn87z44ov89V//NX/+53/O3XffrT4tfc/r9TI9Pc2v/dqv8fnPf549e/asm36ta+w2GMMwqFari4+73S52u7qB9KdUKsWdd97J7t27uemmm/irv/qrxWnVapVAINDD6kRW5vHHH8disfC///u/HD58mHvuuYf5+fnF6erT0m9CoRATExM4nU4mJiZwuVzMzs4uTlefln70D//wD1x77bV85jOfIZVK8bu/+7u0Wq3F6b3s1zpit8Hs2rWLAwcOAHDo0CEmJyd7XJHIhclms/z+7/8+f/RHf8SHPvQhAN72trfxzDPPAHDgwAHe8Y539LJEkRX5zne+wz/+4z+yb98+rrjiCr74xS9y/fXXq09L37rmmmv47//+b0zTZG5ujoWFBX7+539efVr6WiAQwO/3AxAMBmm32+tm+0MDlG8wb9wV89ixY5imyYMPPsi2bdt6XZbIiu3du5cnnniCiYmJxec+97nPsXfvXlqtFhMTE+zduxebzdbDKkUuzJ49e7j//vuxWq18/vOfV5+WvvWXf/mXPPPMM5imyV133cXY2Jj6tPS1arXKZz/7WTKZDK1Wi9tuu42rrrpqXfRrBTsREREREZE+p1MxRURERERE+pyCnYiIiIiISJ9TsBMREREREelzCnYiIiIiIiJ9TsFORERERESkzynYiYhIX2s0GvzTP/3TWc9lMhnuv//+FS9rz549HD9+fFXqOnr0KD/84Q8BuOGGG2g0GquyXBERkXNRsBMRkb6WyWR+JtjF4/ELCnar6d///d955ZVXelqDiIhsHPZeFyAiInIxvvGNb/DKK69w+eWX8wu/8AvUajX+4i/+gj/5kz/hscce4/3vfz/veMc7ePnllwkGg3z5y1/G6/Wed5nlcpnPfe5z5PN5AP70T/+Uyy67jF/+5V9m165dvPbaa0SjUb761a/SarX44z/+Y9LpNMPDw/zwhz/k8ccf51//9V9xOBxceeWVANx///1MTU0B8LWvfY1gMLi2b4yIiGwoOmInIiJ97fbbb2f79u3ceeedTExM8Oijj+JyuRan1+t1brrpJr73ve8xMTHB/v37l1zmN77xDd797nezb98+HnjggcWjf6dPn+ZTn/oU+/fvZ35+nh/96Efs37+fsbExHn30Uf7gD/6AXC5HMpnkt37rt/jIRz7Czp07AfjgBz/Ivn37GB0d5emnn16T90JERDYuHbETEZGBsXXr1p95zm638853vhOAXbt2ceDAgSWXc+zYMX7wgx/wxBNPAFAqlQAIh8MMDw8DMDw8TKPR4Pjx41x//fUAbNu2jUgkcs5lXnXVVQDEYjHq9foK10xEROT8FOxERKSvWa1Wut3u4s8/rd1uc+TIES6//HIOHjzI9u3bl1zmxMQEv/Ebv8FNN91ELpdbvIbPYrH8zLyTk5M8//zz/NIv/RKnTp1aPH3TYrEs1vVWrxUREVktOhVTRET6WjQapdVqnfco2Le+9S0+/OEPk06nufXWW5dc5u23384TTzzBnj17+NjHPsaOHTvect4PfehDTE9P89u//dt89atfXTwN9KqrruI73/kOP/jBD1a+UiIiIitkMU3T7HURIiIia+WGG27giSeeOOu6u9X03HPPUavVuPbaazlx4gQf+9jHePLJJ9ekLRERkbeiUzFFRGRDmZmZ4Z577vmZ59/5znfyyU9+csXL27RpE3/4h3/I1772NdrtNvfdd99qlCkiIrIiOmInIiIiIiLS53SNnYiIiIiISJ9TsBMREREREelzCnYiIiIiIiJ9TsFORERERESkzynYiYiIiIiI9DkFOxERERERkT73/6BoR5pMlSBsAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "printmd('**Descriptive Statistics of Trip Length after Outliers Removed:**')\n",
    "display(subway_trip_mor_clean[['trip_length']].describe())\n",
    "\n",
    "plot_distribution(df=subway_trip_mor_clean, col='trip_length', bins=25)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classify Trips into Three Types\n",
    "\n",
    "We're now familiar with the trip length distribution. Let's classify these trips into three types - short, medium, long distance. By classifying trips (and passengers behind) into different types, we can compute the share of each trip type in all stations, and identify which stations have the largest share of long distance passengers. These stations are the favorable ones to launch ads.\n",
    "\n",
    "<mark>**Share of each Trip Type**</mark>\n",
    "\n",
    "- Short distance trips (passengers) - Trip length less than 20 mins. (accounts for 26%)\n",
    "- Medium distance trips (passengers) - Trip length from 20 to 60 mins. (accounts for 67%)\n",
    "- Long distance trips (passengers) - Trip length longer than 60 mins. (accounts for 7%)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def bin_cut(df, col, bin_range):\n",
    "    target_bin = pd.cut(df[col], bin_range, duplicates='drop') # drop off bins with the same index, incur less bin number\n",
    "\n",
    "    print('Share of each bin:')\n",
    "    bin_share = target_bin.groupby(target_bin).agg({'size': lambda x: x.size,\n",
    "                                                    'share': lambda x: x.size/len(target_bin)})\n",
    "    display(bin_share)\n",
    "\n",
    "    map_class = {}\n",
    "    for i, key in enumerate(sorted(target_bin.unique())):\n",
    "        map_class[key] = i\n",
    "    print('Bin and class label match:')\n",
    "    display(map_class)\n",
    "\n",
    "    target_bin_2 = target_bin.replace(map_class)\n",
    "    target_bin = pd.concat([target_bin, target_bin_2], axis=1)\n",
    "    target_bin.columns = ['{}_bin'.format(col), '{}_class'.format(col)]\n",
    "    \n",
    "    new_df = pd.concat([df, target_bin], axis=1)\n",
    "    return new_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Share of each bin:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>size</th>\n",
       "      <th>share</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>trip_length</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>(-1, 20]</th>\n",
       "      <td>19581</td>\n",
       "      <td>0.26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>(20, 60]</th>\n",
       "      <td>51401</td>\n",
       "      <td>0.67</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>(60, 362]</th>\n",
       "      <td>5171</td>\n",
       "      <td>0.07</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              size  share\n",
       "trip_length              \n",
       "(-1, 20]     19581   0.26\n",
       "(20, 60]     51401   0.67\n",
       "(60, 362]     5171   0.07"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Bin and class label match:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{Interval(-1, 20, closed='right'): 0,\n",
       " Interval(20, 60, closed='right'): 1,\n",
       " Interval(60, 362, closed='right'): 2}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "subway_trip_mor_cl = bin_cut(\n",
    "    df=subway_trip_mor_clean, \n",
    "    col='trip_length', \n",
    "    bin_range=[-1, 20, 60, subway_trip_mor['trip_length'].max()])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Aggregate to Destination Station\n",
    "\n",
    "Now I've classified each trip to short, medium, long distance trip. I can aggregate all trips up to station level, so that to see **the share of each trip type on each station.**\n",
    "\n",
    "To be more precise here, the station in the context is the destination of each trip. What really matters is for each station, how long passengers have traveled to arrive that station, so that we know in which stations we're mostly likely to encounter long distance passengers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Count trip numbers.\n",
    "station_share = subway_trip_mor_cl.groupby(['line_y','stop_y', 'trip_length_class']).count(\n",
    "                    )[['trip_length_bin']].rename(columns={'trip_length_bin':'count'})\n",
    "\n",
    "# long to wide\n",
    "station_share = station_share.unstack(level='trip_length_class')\n",
    "\n",
    "# remove multi-columns\n",
    "cols = station_share.columns.droplevel()\n",
    "cols = ['class_' + str(col) for col in cols]\n",
    "station_share.columns = cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Share of Each Type in Stations:** \n",
       " (class_0: short distance, class_1: medium distance, class_2: long distance)"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>class_0</th>\n",
       "      <th>class_1</th>\n",
       "      <th>class_2</th>\n",
       "      <th>total</th>\n",
       "      <th>class_0_rate</th>\n",
       "      <th>class_1_rate</th>\n",
       "      <th>class_2_rate</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>line_y</th>\n",
       "      <th>stop_y</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">2</th>\n",
       "      <th>海天三路</th>\n",
       "      <td>4.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>8.00</td>\n",
       "      <td>17.00</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>凌空路</th>\n",
       "      <td>1.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0.14</td>\n",
       "      <td>0.43</td>\n",
       "      <td>0.43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">16</th>\n",
       "      <th>书院</th>\n",
       "      <td>3.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>11.00</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>航头东</th>\n",
       "      <td>4.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>11.00</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <th>美兰湖</th>\n",
       "      <td>7.00</td>\n",
       "      <td>37.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>68.00</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.54</td>\n",
       "      <td>0.35</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               class_0  class_1  class_2  total  class_0_rate  class_1_rate  \\\n",
       "line_y stop_y                                                                 \n",
       "2      海天三路       4.00     5.00     8.00  17.00          0.24          0.29   \n",
       "       凌空路        1.00     3.00     3.00   7.00          0.14          0.43   \n",
       "16     书院         3.00     4.00     4.00  11.00          0.27          0.36   \n",
       "       航头东        4.00     3.00     4.00  11.00          0.36          0.27   \n",
       "7      美兰湖        7.00    37.00    24.00  68.00          0.10          0.54   \n",
       "\n",
       "               class_2_rate  \n",
       "line_y stop_y                \n",
       "2      海天三路            0.47  \n",
       "       凌空路             0.43  \n",
       "16     书院              0.36  \n",
       "       航头东             0.36  \n",
       "7      美兰湖             0.35  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate percentage.\n",
    "station_share['total'] = station_share['class_0'] + station_share['class_1'] + station_share['class_2']\n",
    "station_share['class_0_rate'] = station_share['class_0'] / station_share['total']\n",
    "station_share['class_1_rate'] = station_share['class_1'] / station_share['total']\n",
    "station_share['class_2_rate'] = station_share['class_2'] / station_share['total']\n",
    "station_share = station_share.sort_values(by='class_2_rate', ascending=False)\n",
    "\n",
    "printmd('**Share of Each Type in Stations:** \\n (class_0: short distance, class_1: medium distance, class_2: long distance)')\n",
    "display(station_share.head(n=5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Statistics of Distribution\n",
    "\n",
    "From the table below, we can tell that roughly each station has **10%** of passengers traveling long distance to get there, in terms of absolute number, that would be **16** people. But noted, it's computed only on sampled data, which accounts for **4%** of the population. Thus, the absolute number of trips (passengers) must be way higher than the figures what we see here.\n",
    "\n",
    "We can even visualize it in a stacked bar plot, and get a general view of how many there are stations having more-than-average-share long distance passengers. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Descriptive Statistics of Trip Type Share among Stations:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>class_0</th>\n",
       "      <th>class_1</th>\n",
       "      <th>class_2</th>\n",
       "      <th>total</th>\n",
       "      <th>class_0_rate</th>\n",
       "      <th>class_1_rate</th>\n",
       "      <th>class_2_rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>312.00</td>\n",
       "      <td>313.00</td>\n",
       "      <td>311.00</td>\n",
       "      <td>311.00</td>\n",
       "      <td>311.00</td>\n",
       "      <td>311.00</td>\n",
       "      <td>311.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>62.76</td>\n",
       "      <td>164.22</td>\n",
       "      <td>16.63</td>\n",
       "      <td>244.75</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.63</td>\n",
       "      <td>0.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>68.29</td>\n",
       "      <td>196.57</td>\n",
       "      <td>18.57</td>\n",
       "      <td>274.51</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>17.00</td>\n",
       "      <td>39.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>68.00</td>\n",
       "      <td>0.21</td>\n",
       "      <td>0.59</td>\n",
       "      <td>0.05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>39.00</td>\n",
       "      <td>86.00</td>\n",
       "      <td>11.00</td>\n",
       "      <td>143.00</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.66</td>\n",
       "      <td>0.07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>81.25</td>\n",
       "      <td>210.00</td>\n",
       "      <td>20.00</td>\n",
       "      <td>308.00</td>\n",
       "      <td>0.31</td>\n",
       "      <td>0.71</td>\n",
       "      <td>0.13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>453.00</td>\n",
       "      <td>1496.00</td>\n",
       "      <td>208.00</td>\n",
       "      <td>2016.00</td>\n",
       "      <td>0.63</td>\n",
       "      <td>0.85</td>\n",
       "      <td>0.47</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       class_0  class_1  class_2   total  class_0_rate  class_1_rate  \\\n",
       "count   312.00   313.00   311.00  311.00        311.00        311.00   \n",
       "mean     62.76   164.22    16.63  244.75          0.27          0.63   \n",
       "std      68.29   196.57    18.57  274.51          0.09          0.10   \n",
       "min       1.00     3.00     1.00    6.00          0.05          0.24   \n",
       "25%      17.00    39.00     6.00   68.00          0.21          0.59   \n",
       "50%      39.00    86.00    11.00  143.00          0.25          0.66   \n",
       "75%      81.25   210.00    20.00  308.00          0.31          0.71   \n",
       "max     453.00  1496.00   208.00 2016.00          0.63          0.85   \n",
       "\n",
       "       class_2_rate  \n",
       "count        311.00  \n",
       "mean           0.10  \n",
       "std            0.07  \n",
       "min            0.01  \n",
       "25%            0.05  \n",
       "50%            0.07  \n",
       "75%            0.13  \n",
       "max            0.47  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "printmd('**Descriptive Statistics of Trip Type Share among Stations:**')\n",
    "station_share.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot stacked barplot\n",
    "fig, ax = plt.subplots(1,1,figsize=(25,10))\n",
    "\n",
    "_ = station_share.plot(y=['class_2_rate', 'class_1_rate', 'class_0_rate'],\n",
    "                        kind='bar', ax=ax, rot=30, stacked=True,\n",
    "                        colors=['#7c260b','#cec3c6','#014d64'])\n",
    "_ = ax.tick_params(labelbottom=False) # remove xtick labels\n",
    "_ = ax.legend(fontsize='xx-large')\n",
    "_ = ax.set_title('Share of Each Passenger Type over Each Station', fontsize=20)\n",
    "_ = ax.set_xlabel(xlabel='Stations', fontsize=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Suggested List of Stations\n",
    "\n",
    "Finally it comes to answer our question. Which stations are the first choices when planning placing flat leashing ads.\n",
    "\n",
    "Let's try to take **75% percentile** as the filtering criterion, meaning to select only station which has absolute number of long distance passengers more than **20 people** and the share of long distance passengers higher than **13%**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**In total, 20 stations are selected, as follow,**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>class_0</th>\n",
       "      <th>class_1</th>\n",
       "      <th>class_2</th>\n",
       "      <th>total</th>\n",
       "      <th>class_0_rate</th>\n",
       "      <th>class_1_rate</th>\n",
       "      <th>class_2_rate</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>line_y</th>\n",
       "      <th>stop_y</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <th>美兰湖</th>\n",
       "      <td>7.00</td>\n",
       "      <td>37.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>68.00</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.54</td>\n",
       "      <td>0.35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <th>佘山</th>\n",
       "      <td>33.00</td>\n",
       "      <td>46.00</td>\n",
       "      <td>34.00</td>\n",
       "      <td>113.00</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.41</td>\n",
       "      <td>0.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <th>航津路</th>\n",
       "      <td>6.00</td>\n",
       "      <td>42.00</td>\n",
       "      <td>20.00</td>\n",
       "      <td>68.00</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.62</td>\n",
       "      <td>0.29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <th>沈杜公路</th>\n",
       "      <td>21.00</td>\n",
       "      <td>122.00</td>\n",
       "      <td>52.00</td>\n",
       "      <td>195.00</td>\n",
       "      <td>0.11</td>\n",
       "      <td>0.63</td>\n",
       "      <td>0.27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <th>虹桥火车站</th>\n",
       "      <td>123.00</td>\n",
       "      <td>498.00</td>\n",
       "      <td>208.00</td>\n",
       "      <td>829.00</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.60</td>\n",
       "      <td>0.25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <th>惠南</th>\n",
       "      <td>14.00</td>\n",
       "      <td>52.00</td>\n",
       "      <td>21.00</td>\n",
       "      <td>87.00</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.60</td>\n",
       "      <td>0.24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <th>嘉定北</th>\n",
       "      <td>23.00</td>\n",
       "      <td>72.00</td>\n",
       "      <td>29.00</td>\n",
       "      <td>124.00</td>\n",
       "      <td>0.19</td>\n",
       "      <td>0.58</td>\n",
       "      <td>0.23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <th>徐泾东</th>\n",
       "      <td>42.00</td>\n",
       "      <td>129.00</td>\n",
       "      <td>52.00</td>\n",
       "      <td>223.00</td>\n",
       "      <td>0.19</td>\n",
       "      <td>0.58</td>\n",
       "      <td>0.23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <th>顾村公园</th>\n",
       "      <td>48.00</td>\n",
       "      <td>45.00</td>\n",
       "      <td>26.00</td>\n",
       "      <td>119.00</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.38</td>\n",
       "      <td>0.22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <th>松江大学城</th>\n",
       "      <td>39.00</td>\n",
       "      <td>86.00</td>\n",
       "      <td>31.00</td>\n",
       "      <td>156.00</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.55</td>\n",
       "      <td>0.20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <th>东川路</th>\n",
       "      <td>45.00</td>\n",
       "      <td>63.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>132.00</td>\n",
       "      <td>0.34</td>\n",
       "      <td>0.48</td>\n",
       "      <td>0.18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <th>江月路</th>\n",
       "      <td>32.00</td>\n",
       "      <td>111.00</td>\n",
       "      <td>31.00</td>\n",
       "      <td>174.00</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.64</td>\n",
       "      <td>0.18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <th>国权路</th>\n",
       "      <td>30.00</td>\n",
       "      <td>73.00</td>\n",
       "      <td>22.00</td>\n",
       "      <td>125.00</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.58</td>\n",
       "      <td>0.18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <th>芳华路</th>\n",
       "      <td>21.00</td>\n",
       "      <td>89.00</td>\n",
       "      <td>22.00</td>\n",
       "      <td>132.00</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.67</td>\n",
       "      <td>0.17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <th>富锦路</th>\n",
       "      <td>39.00</td>\n",
       "      <td>65.00</td>\n",
       "      <td>20.00</td>\n",
       "      <td>124.00</td>\n",
       "      <td>0.31</td>\n",
       "      <td>0.52</td>\n",
       "      <td>0.16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <th>南翔</th>\n",
       "      <td>53.00</td>\n",
       "      <td>146.00</td>\n",
       "      <td>38.00</td>\n",
       "      <td>237.00</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.62</td>\n",
       "      <td>0.16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <th>颛桥</th>\n",
       "      <td>87.00</td>\n",
       "      <td>71.00</td>\n",
       "      <td>27.00</td>\n",
       "      <td>185.00</td>\n",
       "      <td>0.47</td>\n",
       "      <td>0.38</td>\n",
       "      <td>0.15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <th>广兰路</th>\n",
       "      <td>86.00</td>\n",
       "      <td>334.00</td>\n",
       "      <td>71.00</td>\n",
       "      <td>491.00</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.68</td>\n",
       "      <td>0.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <th>花木路</th>\n",
       "      <td>95.00</td>\n",
       "      <td>253.00</td>\n",
       "      <td>56.00</td>\n",
       "      <td>404.00</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.63</td>\n",
       "      <td>0.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <th>龙柏新村</th>\n",
       "      <td>34.00</td>\n",
       "      <td>112.00</td>\n",
       "      <td>22.00</td>\n",
       "      <td>168.00</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.67</td>\n",
       "      <td>0.13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               class_0  class_1  class_2  total  class_0_rate  class_1_rate  \\\n",
       "line_y stop_y                                                                 \n",
       "7      美兰湖        7.00    37.00    24.00  68.00          0.10          0.54   \n",
       "9      佘山        33.00    46.00    34.00 113.00          0.29          0.41   \n",
       "6      航津路        6.00    42.00    20.00  68.00          0.09          0.62   \n",
       "8      沈杜公路      21.00   122.00    52.00 195.00          0.11          0.63   \n",
       "2      虹桥火车站    123.00   498.00   208.00 829.00          0.15          0.60   \n",
       "16     惠南        14.00    52.00    21.00  87.00          0.16          0.60   \n",
       "11     嘉定北       23.00    72.00    29.00 124.00          0.19          0.58   \n",
       "2      徐泾东       42.00   129.00    52.00 223.00          0.19          0.58   \n",
       "7      顾村公园      48.00    45.00    26.00 119.00          0.40          0.38   \n",
       "9      松江大学城     39.00    86.00    31.00 156.00          0.25          0.55   \n",
       "5      东川路       45.00    63.00    24.00 132.00          0.34          0.48   \n",
       "8      江月路       32.00   111.00    31.00 174.00          0.18          0.64   \n",
       "10     国权路       30.00    73.00    22.00 125.00          0.24          0.58   \n",
       "7      芳华路       21.00    89.00    22.00 132.00          0.16          0.67   \n",
       "1      富锦路       39.00    65.00    20.00 124.00          0.31          0.52   \n",
       "11     南翔        53.00   146.00    38.00 237.00          0.22          0.62   \n",
       "5      颛桥        87.00    71.00    27.00 185.00          0.47          0.38   \n",
       "2      广兰路       86.00   334.00    71.00 491.00          0.18          0.68   \n",
       "7      花木路       95.00   253.00    56.00 404.00          0.24          0.63   \n",
       "10     龙柏新村      34.00   112.00    22.00 168.00          0.20          0.67   \n",
       "\n",
       "               class_2_rate  \n",
       "line_y stop_y                \n",
       "7      美兰湖             0.35  \n",
       "9      佘山              0.30  \n",
       "6      航津路             0.29  \n",
       "8      沈杜公路            0.27  \n",
       "2      虹桥火车站           0.25  \n",
       "16     惠南              0.24  \n",
       "11     嘉定北             0.23  \n",
       "2      徐泾东             0.23  \n",
       "7      顾村公园            0.22  \n",
       "9      松江大学城           0.20  \n",
       "5      东川路             0.18  \n",
       "8      江月路             0.18  \n",
       "10     国权路             0.18  \n",
       "7      芳华路             0.17  \n",
       "1      富锦路             0.16  \n",
       "11     南翔              0.16  \n",
       "5      颛桥              0.15  \n",
       "2      广兰路             0.14  \n",
       "7      花木路             0.14  \n",
       "10     龙柏新村            0.13  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create a threshold to identify long-distance passengers accounts for over 75% pecentile\n",
    "# on both the absolute number of passengers and share of passenger type.\n",
    "\n",
    "station_list = station_share[(station_share['class_2'] >= 20) & (station_share['class_2_rate'] >= 0.13)]\n",
    "printmd('**In total, {} stations are selected, as follow,**'.format(station_list.shape[0]))\n",
    "display(station_list)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Expected Commute Time for each Route\n",
    "\n",
    "<mark>**User Scenario**</mark><br/>\n",
    "\n",
    "Assume that passengers saw the ads at the station, and decide to browse on our house leasing website for flat choices. If we were to add on new features in our website, perhaps say, we could provide information of how much travel time required if passengers choose to live in flats around this region. \n",
    "\n",
    "Put it more bluntly, users may directly select that s/he wants to arrive at the office before *11am*. Our website can instantly come up with a suggest list of possible locations to search for flats. That's definitely something going to impress the users.\n",
    "\n",
    "---\n",
    "\n",
    "<mark>**Brief Summary**</mark><br/>\n",
    "\n",
    "Before I jump right to compute expected trip time, I will first create flag on different time slot, so that I can dig in further and explore if trip time varies when we take subway at different timing. \n",
    "\n",
    "For the scarcity of samples on each route, I will simply create two slots - peak and off-peak hour and explore on it.\n",
    "\n",
    "Now as usual, first, let's have a look at the distribution of trip time difference between two time slots and then compute the expected commute time.\n",
    "\n",
    "The **quick conclusion** here is that it seems the time you take subway between peak and off-peak hour doesn't influence the trip time you might spend.\n",
    "\n",
    "Nonetheless, this is worthy of further investigation to be confirmed. But in my case, the expected commute time just suffices, and no need to segment further on peak and off-peak hours. Finally, in the end of this section, a table of expected commute time will be included as well."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Flag Peak and Off-Peak Hours\n",
    "\n",
    "The computation of peak hours is based on the number of exits of all stations over each hour. See the graph down below, it's clear that there are two timing as peak hours - **8:00 and 18:00**. The rest hours have roughly the same amount of traffic."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use the destination's datetime as the interval.\n",
    "route = subway_trip.copy()\n",
    "route['hour_y'] = route['datetime_y'].apply(lambda x: x.hour)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot distribution.\n",
    "peak = route.groupby(['hour_y']).count(\n",
    "    )[['transit_card_id_y']].rename(columns={'transit_card_id_y':'count'})\n",
    "\n",
    "fix, ax = plt.subplots(1,1,figsize=(20,5))\n",
    "\n",
    "_ = peak.plot(y='count', kind='line', ax=ax, colors=['#014d64'])\n",
    "_ = ax.set_xticks(ticks=np.arange(24))\n",
    "_ = ax.legend(fontsize='xx-large')\n",
    "_ = ax.set_title('Exist Numbers of All Stations over Hour', fontsize=20)\n",
    "_ = ax.set_xlabel(xlabel='Hour', fontsize=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create flag for peak_hour or off-peak hour\n",
    "peak_hour = set([8,18])\n",
    "route['peak_hour'] = route['hour_y'].apply(lambda x: x in peak_hour)\n",
    "\n",
    "# display(route.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Route's Travel Time between Peak and Off-Peak"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Aggregate up to peak and off-peak interval.\n",
    "route_peak = route.groupby(['line_x','stop_x','line_y','stop_y','peak_hour']).agg(\n",
    "    {'trip_length':[('mean', np.mean), ('count', np.size)]})\n",
    "route_peak.columns = [a + '_' + b for a,b in route_peak.columns.values]\n",
    "# display(route_peak.head(n=2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Check if rows contain missing values:\n",
      "\n",
      "line_x                          False\n",
      "stop_x                          False\n",
      "line_y                          False\n",
      "stop_y                          False\n",
      "trip_length_mean_peak_false     False\n",
      "trip_length_mean_peak_true      False\n",
      "trip_length_count_peak_false    False\n",
      "trip_length_count_peak_true     False\n",
      "peak_diff                       False\n",
      "dtype: bool\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "\n",
       "**Preview of route_peak_wide:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>line_x</th>\n",
       "      <th>stop_x</th>\n",
       "      <th>line_y</th>\n",
       "      <th>stop_y</th>\n",
       "      <th>trip_length_mean_peak_false</th>\n",
       "      <th>trip_length_mean_peak_true</th>\n",
       "      <th>trip_length_count_peak_false</th>\n",
       "      <th>trip_length_count_peak_true</th>\n",
       "      <th>peak_diff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>上海南站</td>\n",
       "      <td>8.60</td>\n",
       "      <td>9.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>上海火车站</td>\n",
       "      <td>22.33</td>\n",
       "      <td>26.50</td>\n",
       "      <td>3.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>4.17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>上海马戏城</td>\n",
       "      <td>33.50</td>\n",
       "      <td>32.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>-1.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>中山北路</td>\n",
       "      <td>28.00</td>\n",
       "      <td>26.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>-2.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>共康路</td>\n",
       "      <td>36.00</td>\n",
       "      <td>40.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>4.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  line_x stop_x line_y stop_y  trip_length_mean_peak_false  \\\n",
       "1      1  上海体育馆      1   上海南站                         8.60   \n",
       "2      1  上海体育馆      1  上海火车站                        22.33   \n",
       "3      1  上海体育馆      1  上海马戏城                        33.50   \n",
       "4      1  上海体育馆      1   中山北路                        28.00   \n",
       "5      1  上海体育馆      1    共康路                        36.00   \n",
       "\n",
       "   trip_length_mean_peak_true  trip_length_count_peak_false  \\\n",
       "1                        9.00                          5.00   \n",
       "2                       26.50                          3.00   \n",
       "3                       32.00                          2.00   \n",
       "4                       26.00                          3.00   \n",
       "5                       40.00                          1.00   \n",
       "\n",
       "   trip_length_count_peak_true  peak_diff  \n",
       "1                         1.00       0.40  \n",
       "2                         2.00       4.17  \n",
       "3                         1.00      -1.50  \n",
       "4                         1.00      -2.00  \n",
       "5                         1.00       4.00  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# long to wide\n",
    "route_peak_wide = route_peak.unstack(level='peak_hour')\n",
    "cols = route_peak_wide.columns\n",
    "cols = [str(a).lower() + '_peak_' + str(b).lower() for a,b in cols] \n",
    "route_peak_wide.columns = cols\n",
    "route_peak_wide = route_peak_wide.dropna()\n",
    "route_peak_wide = route_peak_wide.reset_index()\n",
    "route_peak_wide['peak_diff'] = route_peak_wide['trip_length_mean_peak_true'] - route_peak_wide['trip_length_mean_peak_false']\n",
    "print('Check if rows contain missing values:\\n')\n",
    "print(route_peak_wide.isna().any())\n",
    "\n",
    "printmd('\\n**Preview of route_peak_wide:**')\n",
    "display(route_peak_wide[1:].head(n=5))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Distribution of Trip Time Difference on Peak and Off-Peak Hour\n",
    "\n",
    "**First, delete abnormal routes.** There are some cases where passengers get on and get off at the same station. These records will be excluded from dataset. \n",
    "\n",
    "Table below is the statistics of each subway route's commute time. From the table, we can tell that commute time doesn't vary so much between **peak(mean 41 mins)** and **off-peak(mean 40 mins)** hours. Thus, no need to differentiate commute time over different time slot.\n",
    "\n",
    "*Noted that it is derived from sampled dataset, therefore lots of routes have relative few samples to infer its mean travel time*. That's something to be wary of when interpreting this data.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "237 abnormal routes are deleted from dataset.\n"
     ]
    }
   ],
   "source": [
    "# Exclude route where get-on and get-off station are the same.\n",
    "delete_idx = ((route_peak_wide['line_x'] == route_peak_wide['line_y']) & \n",
    "              (route_peak_wide['stop_x'] == route_peak_wide['stop_y']))\n",
    "print('{} abnormal routes are deleted from dataset.'.format(sum(delete_idx)))\n",
    "\n",
    "route_peak_wide_clean = route_peak_wide[~delete_idx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Descriptive Statistics of Routes on Peak and Off-Peak:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>trip_length_mean_peak_false</th>\n",
       "      <th>trip_length_mean_peak_true</th>\n",
       "      <th>trip_length_count_peak_false</th>\n",
       "      <th>trip_length_count_peak_true</th>\n",
       "      <th>peak_diff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>20157.00</td>\n",
       "      <td>20157.00</td>\n",
       "      <td>20157.00</td>\n",
       "      <td>20157.00</td>\n",
       "      <td>20157.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>41.68</td>\n",
       "      <td>40.01</td>\n",
       "      <td>4.80</td>\n",
       "      <td>2.30</td>\n",
       "      <td>-1.67</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>36.82</td>\n",
       "      <td>39.59</td>\n",
       "      <td>6.61</td>\n",
       "      <td>2.83</td>\n",
       "      <td>46.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>-889.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>24.80</td>\n",
       "      <td>23.60</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>-3.50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>36.00</td>\n",
       "      <td>35.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>-0.67</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>50.00</td>\n",
       "      <td>48.50</td>\n",
       "      <td>5.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>2.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>916.00</td>\n",
       "      <td>781.00</td>\n",
       "      <td>118.00</td>\n",
       "      <td>93.00</td>\n",
       "      <td>771.50</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       trip_length_mean_peak_false  trip_length_mean_peak_true  \\\n",
       "count                     20157.00                    20157.00   \n",
       "mean                         41.68                       40.01   \n",
       "std                          36.82                       39.59   \n",
       "min                           1.00                        1.00   \n",
       "25%                          24.80                       23.60   \n",
       "50%                          36.00                       35.00   \n",
       "75%                          50.00                       48.50   \n",
       "max                         916.00                      781.00   \n",
       "\n",
       "       trip_length_count_peak_false  trip_length_count_peak_true  peak_diff  \n",
       "count                      20157.00                     20157.00   20157.00  \n",
       "mean                           4.80                         2.30      -1.67  \n",
       "std                            6.61                         2.83      46.79  \n",
       "min                            1.00                         1.00    -889.50  \n",
       "25%                            1.00                         1.00      -3.50  \n",
       "50%                            3.00                         1.00      -0.67  \n",
       "75%                            5.00                         3.00       2.00  \n",
       "max                          118.00                        93.00     771.50  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "printmd('**Descriptive Statistics of Routes on Peak and Off-Peak:**')\n",
    "display(route_peak_wide_clean.describe())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_distribution(df=route_peak_wide_clean, col='peak_diff', bins=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create Table of Mean Commute Time for Each Route\n",
    "\n",
    "<mark>**User Scenario**</mark><br/>\n",
    "\n",
    "Below is a preview of final dataset on **mean commute time for each subway route**. It can be used as reference when users are searching for flats within acceptable commute time.\n",
    "\n",
    "<mark>**Feature Engineering**</mark><br/>\n",
    "\n",
    "Actually, if we aim at building up a model, **mean commute time of each route** can also serve as a very **informative feature**. Let's say we want to make prediction of the flat's rent, the mean time to commute from the flat to major business district absolutely plays a role."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "**Preview of Table for Mean Commute Time on each Route:**"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>line_x</th>\n",
       "      <th>stop_x</th>\n",
       "      <th>line_y</th>\n",
       "      <th>stop_y</th>\n",
       "      <th>trip_length_mean</th>\n",
       "      <th>trip_length_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>11.40</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>上海南站</td>\n",
       "      <td>8.67</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>上海火车站</td>\n",
       "      <td>24.00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>上海马戏城</td>\n",
       "      <td>33.00</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>上海体育馆</td>\n",
       "      <td>1</td>\n",
       "      <td>中山北路</td>\n",
       "      <td>27.50</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  line_x stop_x line_y stop_y  trip_length_mean  trip_length_count\n",
       "0      1  上海体育馆      1  上海体育馆             11.40                  5\n",
       "1      1  上海体育馆      1   上海南站              8.67                  6\n",
       "2      1  上海体育馆      1  上海火车站             24.00                  5\n",
       "3      1  上海体育馆      1  上海马戏城             33.00                  3\n",
       "4      1  上海体育馆      1   中山北路             27.50                  4"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create table of mean commute time for each route.\n",
    "route_mean = route.groupby(['line_x','stop_x','line_y','stop_y']).agg(\n",
    "    {'trip_length':[('mean', np.mean), ('count', np.size)]})\n",
    "route_mean.columns = [a + '_' + b for a,b in route_mean.columns.values]\n",
    "\n",
    "printmd('**Preview of Table for Mean Commute Time on each Route:**')\n",
    "route_mean = route_mean.reset_index()\n",
    "display(route_mean.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descriptive Statistics of Routes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>trip_length_mean</th>\n",
       "      <th>trip_length_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>52581.00</td>\n",
       "      <td>52581.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>50.86</td>\n",
       "      <td>3.84</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>46.15</td>\n",
       "      <td>6.24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>30.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>43.50</td>\n",
       "      <td>2.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>61.00</td>\n",
       "      <td>4.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>977.00</td>\n",
       "      <td>180.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       trip_length_mean  trip_length_count\n",
       "count          52581.00           52581.00\n",
       "mean              50.86               3.84\n",
       "std               46.15               6.24\n",
       "min                0.00               1.00\n",
       "25%               30.00               1.00\n",
       "50%               43.50               2.00\n",
       "75%               61.00               4.00\n",
       "max              977.00             180.00"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print('Descriptive Statistics of Routes:')\n",
    "display(route_mean.describe())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Final Summary on Application Scenario out of this Analysis\n",
    "\n",
    "Now, let's sum up of what we've known from this data exploration."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<mark>**1. In Terms of Ads Deployemnt**</mark><br/>\n",
    "\n",
    "In the way of assisting business expansion, my idea is to come up with suggest list of stations to launch ads. Indeed, there're a few stations worth to be considered with its high share of long distance passengers and the traffic volume. \n",
    "\n",
    "<mark>**2. In Terms of Feature Add-On for Website**</mark><br/>\n",
    "\n",
    "What's more better would be in the case, where assuming the passengers saw the ads at the station, and then decide to browse our website. We can not only provide the rent of each flat to them, but also the expected travel time from the your chosen flat to the office. These information all come from the table of mean commute time for each route.\n",
    "\n",
    "<mark>**3. Reminder of the Suggested Timing to Depart from Home**</mark><br/>\n",
    "\n",
    "If there is a huge impact on when you depart from home for work. Then our website can inform the user that the expected time required to commute. So that we can suggest the best timing to leave for the office."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Empower with More Data\n",
    "\n",
    "<mark>**User Scenario**</mark><br/>\n",
    "\n",
    "All the discussion so far is focused on how to identify best marketing channel, how to enhance the experience of flat seekers. But the flat leasing company can also provide better service for the **flat posters (landlords)**. One example is to provide proposed rent to new landlords. With proper posted rent, the landlord can lease out the flat more quickly and this in turn, can attract more landlords to post their listing on the website. This will attract more flat seekers again. All these form up a reinforced cycle.\n",
    "\n",
    "<mark>**Suggested Data**</mark><br/>\n",
    "\n",
    "In order to accomplish the goal, useful data are in need including, **the posted rent, flat location, size, and equipped amenities etc**. With all these data at hand, we can train a prediction model for any flat. Even on newly developed region, we can assess the flat's price and rent even under the limit of few samples. Following section will sketch out the specification of the predictive model in detail."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Proposed Model\n",
    "\n",
    "In the case of predicting rent (numerical variable), there are a lot of models to be chosen from. Common ones include **Regression, Lasso Regression, Decision Tree Regression, and more advanced Neural Network**. Here I will just use the general form of regression to display the idea of how we can build up the predictive model for rent.\n",
    "\n",
    "$$y = \\alpha x_1 + \\beta x_2 + \\gamma x_3 \\dots$$\n",
    "\n",
    "<mark>**Target Variable:**</mark>\n",
    "- **`proposed rent`** \n",
    "\n",
    "<mark>**Features:**</mark>\n",
    "\n",
    "<span style='line-height:2.0'>\n",
    "\n",
    "- **`Flat Attributes`**\n",
    "    - Can be accessed from website's flat posting data.\n",
    "    - Such as 1) flat size, 2) amenities, 3) number of rooms 4) age of flat.\n",
    "- **`Neighborhood`**\n",
    "    - Can be measured from flat's location on map.\n",
    "    - Such as 1) distance to supermarket, 2) distance to public transportation. \n",
    "- **`Number of Available Flats Nearby`**\n",
    "    - Can be accessed from website's flat posting data.\n",
    "    - Such as 1) number of flat posting in the same neighborhood.\n",
    "- **`Commute Time`**\n",
    "    - Can be derived from this traffic dataset.\n",
    "    - Such as 1) mean commute time to major business district.\n",
    "- **`Population of Neighborhood`**\n",
    "    - Can be estimated from this traffic dataset.\n",
    "    - Such as 1) number of residents, either with or without permanent address in that region.\n",
    "\n",
    "</span>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Other Application from this Dataset\n",
    "\n",
    "The dataset we have at hand can actually be applied to other areas as well. Here I will just name two other potential applications leveraging this dataset. One, **metro traffic control** and, the other is **passenger's profiling**.\n",
    "\n",
    "<mark>**Metro Traffic Control**</mark><br/>\n",
    "\n",
    "Apart from the absolute traffic volume, the **change rate** of traffic also challenges the ability of metro company to cope with the sudden spike on demand. As we know the **capacity** of each station should be designed to the **peak demand** instead of **mean traffic volume**. Hence, identifying which station has the highest change rate of traffic demand will enable the authority to take proper measures on re-designing the traffic route within these stations and thus, avoiding the case of overcrowds and human stampedes.\n",
    "\n",
    "<mark>**Passenger's Profiling**</mark><br/>\n",
    "\n",
    "For long, we've always wanted to build up comprehensive consumer's profiling so that company can provide more suitable service to the end users. In previous attempt, many companies have already build up their consumer profiling based on **their demography, their purchasing behavior, their financial conditions** etc, but few of them ever included the information of **consumer's commute behavior**, or travel path. By covering the travel path of each consumer, we can predict their next location, and push relevant information to them in advance. This is something very technical challenging but possess huge business opportunities."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reference\n",
    "\n",
    "- [tqdm_notebook](https://github.com/tqdm/tqdm/issues/375)\n",
    "- [tqdm not showing bar](https://stackoverflow.com/questions/48935907/tqdm-not-showing-bar)\n",
    "- [解决macOS下matplotlib无法显示中文的问题](https://www.jianshu.com/p/8ed59ac76c06)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "258px"
   },
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}