Every day, around two million people use public transport in Switzerland. Despite the presence of various types of flat rate abonnements, there is still a substantial number of people buying single fair tickets on a regular basis. Most of these tickets are still bought at a vending machine, however the number is rapidly decreasing in favor of more convenient distribution channels - smartphones.
The goal of this project is to design, build and evaluate prediction models for recognizing human activities, such as “Riding a bus”, “Walking”, in the context of a user traveling with a mobile phone. The results can further be used to implement an automatic ticket-buying system for public transport. The project will involve developing an application that will collect data from various mobile device sensors, such as accelerometer and gyroscope, as well as performing feature extraction to extract meaningful values out of raw signals. The extracted features are then going to be used to build a supervised classifier that should correctly predict activities for the new data samples.