Missing the Apex: Predicting F1 Races with Python
summary
Claim: Python code reproduces a prior paper then improves it by expanding seasons and including every driver. Implementation uses requests BeautifulSoup pandas to scrape formula1.com, handles canceled sessions and sprints, and saves pivoted CSVs covering 2006–2023 with ~7,800 driver-race entries. Analysis runs contingency coefficients, Spearman rho and an ordinal logistic model and finds qualification is the dominant predictor (contingency 0.63→0.77; qualification beta ≈4× P3).
AI-written summary · about
chapters (11)
- 0:00F1 and Python project intro
- 3:35What is Formula 1
- 8:56Motivation for paper
- 11:01Scraping F1 data
- 15:36Edge cases in scraping
- 17:22Cleaning dataset
- 20:32Baseline for analysis
- 22:18Session correlations with results
- 24:46Engine eras comparison
- 26:33Ordinal logistic regression
- 28:42Next steps for F1 predictions
related talks
39mPredicting sports winners using data analytics with pandas and scikit-learn
41mFootball's data analytics in python
01h 42mRecreating, Understanding, and Visualizing FiveThirtyEight's...
51mFive Ounce Racer - Revving Up Pinewood Derby Software in Python
27mFrom Data to Insights in Minutes: Accelerating Predictive Modelling (in Python) with AutoML
42mFlying Python at 40,000 feet
33mWhy Python Has Taken Over Finance
42mRobust Automated Forecasting in Python and R