I contributed to this open-source reference book by turning machine-learning metrics into concise explanations and visual references. My process started with understanding the mathematics and studying how each metric behaves, then working with the book’s author to decide what needed to be explained visually and what could be left to the reader.
The Little Book of ML Metrics
Route optimisation for waste management — an Omdena challenge, Germany
I like projects where the data has somewhere to go in the real world. Here I worked with geospatial data to model waste-management centres across Germany and build an interface for optimising routes between them: cleaning large geographic datasets, working with OpenStreetMap and QGIS, and connecting the results to the OpenRouteService API.
PythonQGISOSM APIfoliumStreamlit
The project’s folium map of landfill sites across Germany, from the Streamlit app