It's a challenge for the Mckinsey during the Hack UPC 2019. It was great learning more things about data manipulation and traffic accidents.

You can find the relevant files in my Google Drive folder.

The submission file may be a different format therefore it might not be best for validating the model efficiency. It is because I've merged the test_csv file with the vehicles.csv file, therefore there are more than 1 entries per accident id and the predictions for the different entries might be different.

I had another solution using only accidents.csv which was of the correct format, but I decided to include the vehicles.csv in order to include more features.

Built With

Share this project: