Statistical analysis of incident probability and causes on RATP metro lines

Trafic perturbé...again?! This library and this notebook provides a statistical point of view for all these daily incidents occurred on the Paris RATP metro lines.

Using tweets coming from official RATP Twitter accounts (@Ligne1_RATP for line 1 for example), we will see

1. What is the probability of encountering some operational incidents on a particular line?
2. Which lines are more likely to cause everybody unhappy?
3. What are the main causes of these problems?
4. Should I be considered lucky if I never take metros during rush hours?
5. Are there less problems during weekends?
6. Instead of going on vacation, is there any reason to be happy if I still work in August?

Some examples in the year 2018

The fist two examples below refer to the RATP metro line 2, which I take everyday for work.

I should leave work before 17:00 on Wednesday (or after 21:00)

In the figure below you can see the probability of catching an operational incident (trafic perturbé, interrompu...) at a given hour (in fact in the next following hour) and a specific weekday. The maximum value (nearly 9%) can be found on Wednesday at the evening rush hours.

It's us the main responsible for these incidents

Nearly half (48%) of the incidents come from us. Also, 9% is due to unattended bags...

Line 13 should be jealous of line 4

Every Parisian knows that the line 13 is bad. But in the year 2018 it was beaten by the line 4, which records an average probability of incidents larger than 4%.

License

The code is licensed under the terms of the MIT license. The analysis results (text, figures, tables) are licensed under a Creative Commons Attribution 4.0 International License.

Author

Tianyi Li (tianyikillua@gmail.com)