This is repository code for Data Science Weekends competition to predict traffic jam in West Java using Waze Dataset.
For the kaggle competition link and dataset, you can click here
We are in top 13% in private leaderboard 12/96. Team's name : Lulus2021
After did late submission, we could improve our achievement to 5/96 leaderboard.
Our kaggle account:
Kaggle Notebook click here
To predict the result we use various method listed below:
- Catboost
- LGBM
- NODE
- MLP using Pytorch
- FASTAI
- TABNET
- Ensemble Various Method
West Java is one of the largest Province in Indonesia. Over time, West Java continues to develop rapidly. As West Java shares the same issues with the other big Provinces, West Java commences facing traffic jams in several cities. In fact, according to the latest Report by the Jabar Digital Service (JDS) in 2020, Bandung is in the 1st position of the most congested cities, followed by Bekasi and Bogor.
In this event, Data Science Indonesia (DSI), which is the largest Data Science community in Indonesia is recently active at solving problems related to data technological developments. DSI brings Data Science Weekend (DSW), the biggest event in DSI consists of seminars, workshops, talk shows, and competitions in the field of data. DSW collaborated with JDS to see and investigate further problems by inviting all data enthusiasts in Indonesia to participate in building Machine Learning-based models to predict which sectors will have high reports in certain hours. The results of these predictions will be used and directed as input and consideration for JDS and local government in West Java for making policies related to this traffic issue.
Using traffic data provided by the JDS, predict high jam reports in certain sectors at certain dates and hours of a few cities in West Java.
