For the Visualize section, currently we do a random sampling of points, and pull them along with their vectors. Afterwards we process them all in the browser, depending on their dimensionality ...