What does MobilitApp do?

MobilitApp is a citizen science project, where citizens collaborate by generating mobility data with which we develop a prediction model for the mode of transport used. The aim is to help the entities in charge of the city's public transport service to analyze the citizens' mobility habits in order to improve urban mobility and make it more sustainable. MobilitApp periodically collects data from 3 mobile sensors: accelerometer, magnetometer and gyroscope. We do not use GPS to predict the means of transport, this way we save battery life and do not read the person's geolocation constantly.

The information obtained is processed to train a machine learning model capable of subsequently predicting the means of transportation used by citizens. Currently, MobilitApp recognizes these means of transport with more than 88% accuracy: bicycle, electric bicycle, electric scooter, bus, underground, train, tram, motorcycle, car, running and walking.