Smarter-Mobility-Data-Challenge

Author
Arthur Satouf

(Link to the competition) https://codalab.lisn.upsaclay.fr/competitions/7192

Getting started

Below is the structure and scripts used in the challange:

├── Arthur SATOUF.pdf
├── README.md
├── data_to_use
│   ├── data_ewm1.2.csv
│   ├── remCharEWM4.csv
│   ├── test.csv
│   ├── train.csv
│   └── train_onlyNext4EWM4andBack_EWM_remChar.csv
├── main.py
├── notebook
│   ├── area_catboost.ipynb
│   ├── cleaning.ipynb
│   ├── global_catboost.ipynb
│   ├── station_catboost.ipynb
│   └── visualization.ipynb
├── sample_result_submission
│   ├── area.csv
│   ├── global.csv
│   └── station.csv
└── sample_result_submission.zip

• Arthur SATOUF.pdf project report.
• data_to_use - data used to build the models and to forecast
  • train_onlyNext4EWM4andBack_EWM_remChar.csv data set input for Station.
  • data_ewm1.2.csv data set input for Area.
  • remCharEWM4.csv data set inout for Global.
  • train.csv (The basic) it been used as a source to create the above data sets and to get validation index to imporove model for Station by using it as valaidation set.
  • test.csv to be forecasted and submitted.
• main.py to model and build sample_result_submission using CatBoost.
• notebook a bunch of note-bookS to build each model and cleaning.ipynb is used to clean and prepoessing the data to build the input data sets data_to_use
• sample_result_submission my submission.
• sample_result_submission.zip submission en ZIP.

How to run the code

• Clone the project.
• Open Terminal.
• Insert python main.py and wait until it finish to gererat data please note that it would take (15 to 20 min).

The result of the competition