We will introduce tools for visualizing, extracting features and feature engineering and explain the logic and steps needed in order to create them. The techniques will be presented in the context of a simple machine learning model for predicting taxi tariffs and illustrate how we can improve it with these new features.
Participants will be able to learn:
- Visualise spatial data for exploratory data analysis
- Calculate distances
- Extract routes, route distances and estimated durations
- Extract city features based on open source tools, e.g. Open Street Maps
- Data processing performance
- Spatial indices in Python
- Ideas for traffic estimation
- Approaches and potentials for GPS analysis and how they can help you in terms of feature engineering.
We will utilise cloud-based Jupyter notebooks and participants will be able to take the code with them so they can apply for their next project.