As part of a previous project I worked in, I was handed 4 .txt files with GPS readings pertaining to 4 different users. I was asked to analyse them, using whatever means I wanted, in order to learn something useful.
I could have used statistics or machine learning approaches; but having no a-priori knowledge about the data, I had no hypotheses to test or predictions to make. So I just followed a generative approach, exploring the data step-by-step using visualizations, and refining the direction of the analysis as I moved forward. Here is the resulting Jupyter Notebook.
Note: The charts are interactive.