General Instructions

Generating API token

This project uses openrouteservice API to plot maps and routes. So the following steps are necessary at first:

1. Sign up on openrouteservice.org to generate an API token;
2. Create a .env file with the following content: TOKEN=XXXXXXXXXXXXXXX, where XXXXXXXXXXXXXXX is the token generated in the step before;
3. Install the lib: pip install safety-road-mapping
4. Get the road accidents according the instructions bellow: Getting accident road data
5. See the documentation on: safety documentation
6. See the examples on how to use the lib on: how-to-use.ipynb

Getting accident road data

• The accidents data used were extracted from the Polícia Rodoviária Federal website.
• The notebook get_data.ipynb is responsible to download and extract the data used.
• If you want to directly download the files you can click here.
• To consolidate all the accidents by year on an unique .csv file use the notebook consolidate_data.ipynb.
• For a simple explore data analysis from the data look at eda_accidents.ipynb

Adding São Paulo data

• Access the infosiga website
• Download the xlsx file with fatal accidents: Fatal accidents
• Download the csv file with non fatal accidents: Non-fatal accidents
• Use the notebook treat_data_from_SP.ipynb to transform and concatenate the data generating a final file that will be used as input for the safety map code.

Other possible steps on the project

• The accidents data used comes just from road federal police source, so there are some routes that don't receive score because they are state highways.
• Create some unit tests.
• Use Renaest data (currently only SP has geolocation data).
• Include other variables in the score calculation.
• Include driver behaviors data.
• Implement an algorithm to predict accidents.
• The routes subsections are not connected, once they are plotted individually in the map. Visually it can be interesting to connect them. (Is it possible or necessary?).