The usual ways of teaching machine learning involve static visualizations are do not give a good sense of the process of the algorithm. We sought to make it clearer and more engaging.
We created an interactive visualization that is available at https://samueldmeyer.github.io/random-forest-viz/. It uses animation and user interaction to improve learning outcomes. We presented a poster at InfoVis 2017: pdf
I led work on the visualization, while my partner led machine learning analysis.
d3.js, Python sklearn