Thanks for this nice exercise; below a few suggestions:
- It seems main.py as mentioned in the README.md is missing in the repository.
- Would be good to clarify if and if so, how the data is normalized wrt to (at least)
- the number of people using a means of transportation and
- the amount of time people are exposed in the public space while using a means of transportation.
Ideally, you would add a baseline probability matrix between any pair of means of transportation.
Also, it would be good to add an explicit "threats to validity" or "limitation" section to the README, because a catchy data visualization that comes without these is actually a problematic form of Data Science.