Disentangle a pedestrian trajectory analysis tool - How to convert a spaghetti C/C++ tool to a Python library

One of the most crucial parts of pedestrian dynamics research is the analysis of trajectory data. The development of the most used tool in the community started over a decade ago in our institute. In this period, the code base grew steadily, but without focusing on future maintenance. Henceforth, the tool got more challenging to use, from installation to analyzing data. Therefore, we decided on a complete rebuild of the software to make it easier to develop and, more importantly, to use and contribute to, even for inexperienced developers (as most scientists are).
This talk shows how we converted a spaghetti C/C++ tool to a Python library. Not only focusing on good software engineering practices but also taking user feedback into account.