yupi: Generation, Tracking and Analysis of Trajectory data in Python

The study of trajectories produced by the motion of particles, objects or animals is often a core task in many research fields such as biology or robotics. There are many challenges in the process, from the building of trajectories from raw sensor data (images, inertial measurement data, etc) to the recognition of key statistical patterns that may emerge in groups of trajectories. This work introduces a software library that addresses the problem as a whole, covering a full pipeline to process raw data and produce the analytics typically demanded by scientific reports. Unlike other trajectory analysis software, our library focuses on the subject in a quite abstract way, allowing its usage across several fields. We validate the software by the reproduction of key results associated to different original research articles, providing an example script in each case. Our aim is that the present software provides researchers with limited experience in programming or computer vision with an easy-to-handle toolbox for approaching trajectory data.