Why are there not more handy tools out there which help people to understand statistical methods? This was the question I asked myself when creating this notebook on linear discriminant analysis (LDA): Interactive LDA plot / Frederik H. | Observable (answer: because it’s really hard to build them 
)
LDA is a classification technique, which separates data into two (or more) classes. In the notebook you can click to add and remove data points, drag them around and create new populations by controlling the parameters.
Find all the details in my toward data science article:
The notebook can serve to develop all sorts of plots and visualise other statistical methods. Quadratic discriminant analysis would be a natural next step (or Naive Bayes). If anyone would like to collaborate on building another plot, let me know!