← Back to work

Data science · Predictive modeling

F1 Race Analysis

What can we learn about race outcomes before the lights go out?

I use historical Formula 1 data to explore pit stops and predict race outcomes. The prediction models use information available before a race, such as qualifying and recent form.

View repository ↗
Yearly average pit-stop duration by team, exported from the F1 analysis notebook.
Yearly average pit-stop duration by team, exported from the F1 analysis notebook.

About the project

What it is

This project uses historical Formula 1 data to study race performance from two angles. One part looks at pit-stop patterns and team differences. The other builds models that estimate finishing position and podium chances before a race starts.

I kept pre-race prediction separate from the pit-stop analysis so information from the race itself does not leak into the predictions. The work includes reusable data preparation and model pipelines.