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

The study of trajectories resulting from the motion of particles, objects or animals is often the core task in many research fields such as biology or robotics. The challenges in the process extend from how to get a trajectory from raw sensor data (e.g., when tracking) to what kind of statistical tools should be used for modeling or making inferences about populations. This work introduces a software library that addresses the problem as a whole. It contains, for instance, a robust tracking module aiming to make data acquisition handy. Furthermore, it provides a statistical kit for analyzing trajectories, namely, correlation functions, spectral density, parameter estimation, filters, stochastic models to fit against simulations (e.g., the classical Langevin model), among others. Unlike other trajectory analysis software, this library does not make assumptions about the nature of trajectories (e.g., those from GPS), which facilitates its usage across different disciplines. We validated the software by reproducing key results of different original research articles. An example script in each case is presented. We aim to provide researchers with limited experience in programming or computer vision with an easy-to-handle toolbox to manipulate trajectory data.