Sweet Lift Taxi company (fictive company) has collected historical data on taxi orders at airports. To attract more drivers during peak hours, in this project a model should be built to predict the amount of taxi orders for the next hour with a RMSE lower than 48.
Therefore the company supplied a timeseries of rides in the timeframe of 2018-03-01 to 2018-08-31
Within the project the following steps have been carried out:
- Importing and inspecting the data
- Data preprocessing (resampling...)
- In depth analysis of the data
- Feature generation
- Model building, training and evaluation
- Hyperparameter tuning
An Random Forest Regressor with a RMSE of 45.2 has been developed, which exaggerates the requirements of the client