Data analytics and forecasting of bike usage in Bremen, Germany
In this project the values of different bicylce counting stations in the city of Bremen, Germany are analyzed. The goals are:
- Visualize datasets and get a deeper understanding of the presented data
- Answer general questions about the data like:
- Is the bike traffic increasing/decreasing or approximately constant over time?
- Did the COVID pandemic have an influence on the bike usage?
- What are the most used bike ways and why?
- Train a machine learning model, to accurately forecast the bike traffic for a given day. For this different ML models should be used and compared to a baseline model.
The data ranges from 01.01.2013 to 31.12.2022. In addtition to the data of the bicycle counting stations, further data (weather data, geo-locations, national holidays, school vacations) are added.
This project contains 3 notebooks:
- 1_Data_Prep
- 2_EDA
- 3_Imputing
- 4_Model_Training