Climate change is the most pressing challenge humanity is facing. Raising awareness about the environmental impact of our mobility is the first step towards a more sustainable mobility. In this talk I will explain how we apply machine learning algorithms on large-scale Telecom data (around 1M samples per second) in order to quantify the carbon footprint of mobility at the scale of a country. I will then take the example of Switzerland and show how we are able to describe the collective mobility and estimate the daily carbon footprint associated with each canton.