I am a maker of digital products currently living in Amsterdam.
As a data scientist with originally an interaction design background,
I enjoy working with machine learning data problems that require human input.

Genetics Algorithm + Three.js Experiment

I've had some fun playing around with the Genetics algorithm: the fittest nodes of a population evolve with a chance of random mutation, while the unfittest nodes are eliminated. The fitness function for the visualisation below is to maximise the sum of values for all nodes. Each node evolutes in three dimensions: x,y,z, with each dimension having 100 genes. A gene can be either 0 or 1. Therefore, the maximum fitness a node can have is 300.

You can find the code for the Genetics algorithm and the visualisation on Github.