- Kamil Barszczak
- Hanna Jarlaczyńska
- Konrad Pawlik
- 0 - Download Data - Downloading required 2018-taxi-trip-data (We advise to download dataset directly from source link and putting .csv file into a resources directory, as it is faster solution)
- 1 - Data Processing - Renaming columns, getting rid of redundant ones and transforming day of week/time values into ML-ready.
- 2 - Data Visualization - Visualizing number of pickups in each taxi zone.
- 3 - Road Lenghts - Calculating sum of road lengths in each taxi zone.
- 4 - Merge Datasets - Merging taxi data with road lenghts and calculating probability of pickup/dropoff for each zone, day of week and hour.
- 5 - ML Models - training SGDRegressor models to predict a number of passengers and XGBRegressor to predict a probability of pickup/dropoff.
- notebooks - the catalog for jupyter lab scripts that are used to prepare the final dataset.
- processed_data - the catalog is used to store the final processed dataset.
- resource - the catalog contains data used to get the final dataset. For example taxi zones in NY city.
Before running the scripts make sure you have installed:
- xgboost
- pyrosm
- geopandas
- ipyleaflet