Since summer 2020, I've been a participant at Numerai, a hedgefund that runs a daily machine-learning competition and pays out cyrpto rewards for accurate predictions. My current training pipeline, which tunes an ensemble of LightGBM and Keras models after running a boruta feature selection algorithm, can be found on GitHub. With the help of numerai-cli, my automated Python submission pipelines run in Docker containers on AWS nodes.

Data Visualisation

Wealth and population density

The Python code can be found on GitHub. Data sources: Credit Suisse (wealth), European Commission Joint Research Centre (population density), and The World Bank (population).

Top 1% wealth share in 2022