My solution ranked 11th out of 2216 on the Recruit Restaurant Visitor Forecasting Kaggle competition.
The Kaggle data can be found here
The weather data can be found here
The get_rwm.py script calculates the rolling weighted mean for training data,
The get_weather_station.py script finds the station within 20kms which have a high coverage for rain and temperature data for each restaurant.
The prophet.py script uses prophet to generate prediction.
The submission.py script takes data generated by the above 3 scripts as part of the input data. It outputs my final solution.