High-dimensional data, such as the output of neural networks or many biomedical data, cannot be directly visualised as humans can only handle up to three dimensions. A clever insight by Nicola Pezzotti allows the real-time visualisation of similarities in these data.
Based on these similarities, his algorithm calculates an attractive force between the various data points. The big breakthrough, however, was to replace the repulsive force between these points with fields. Kinda like what Einstein did with gravity.
Implementing these fields as textured triangles to be handled by a graphical processor made the algorithm about one hundred times faster. Researchers at the LUMC already used it to discover new cell types., but you can also use it to delve into social media filter bubbles. Or perhaps you would like to play with soccer statistics?
Read more about this insightful algorithm here.