Urban data analysis of vehicle emissions.

Urban emissions represent a major antagonist to cities' livability. In this work, we investigate vehicle emissions using both spatial and non-spatial models to study their relationships with a variety of urban features. We compare such models in terms of performance and interpretation, showing not only that a spatial model has better performances, but also that it is very powerful in explaining the roles played by the predictors in different areas of the city.

The code used for the experiments is organized in .ipynb files in the notebook folder.

# Create conda environment, install dependencies on it and activate it
conda env create --name urbem --file environment.yml
conda activate urbem

# Setup pre-commit and pre-push hooks
pre-commit install -t pre-commit
pre-commit install -t pre-push

To add new dependencies or to update existing ones:

1. Add the name (and version if needed) to the list of dependencies in environment.yml
2. run conda env update --name urbem --file environment.yml --prune
3. Update the file environment.lock.yml by running conda env export > environment.lock.yml

This package was created with the BSCCNS/cookiecutter-data-science project template.