Skip to content

Latest commit

 

History

History
47 lines (32 loc) · 1.84 KB

README.md

File metadata and controls

47 lines (32 loc) · 1.84 KB

Price Prediction for EdTech company

  • Price is a prominent factor in influencing the buying decisions of the people. Price optimization has become crucial for institutes to drive student and profitability.
  • It also plays a major role in churn rate reduction of student.
  • Price optimization reduces the manual work and minimizes the possibility of any human errors.
  • Competitor Price behavior analysis is done which can assist Institutes for making better pricing decisions.

Data Collection and Business Understanding

  • First understanding about business problem.
  • Listing out the various factors affecting the price prediction.
  • Selecting important features for meeting the business target.
  • Web scrapping through various Ed-Tech websites and enquiry.
  • Handled noise data and generated random data.
  • Considering all the necessary details and storing in the database.

Tech Stack

Tech Used: Vs Code for working on models, Jupyter Notebook, Postgresql, Python, html, css

Libraries Used: Pandas for Data Manipulation, matplotlib and seaborn for data visualizaiton, make_pipe for making pipeline, column_transformer for encoding data before training sklearn for data preprocessing and model building, Flask for web application, and heroku for deployment.

EDA and Data Preprocessing

Not much Preprocessing was needed, droped irrelevant columns

Model building

In order to get the best accuracy following models were used:

1. Random Forest Regressor
2. Linear Regression
3. Support Vector Regressor 
4. KNeighbours Regressor

Out of which KNeighbours Regressor performed well with test accuracy of 99% and train accuracy of 98% and pickled for model deployment.

Model deployment

Model was deployed on Heroku.

Deployment link:- https://price-prediction-for-courses.herokuapp.com/