Fast followers? Speed contagion: Assessing the Impact of the Montreal F1 Grand Prix on High-Speed Collision Rates (2000–2022)
This project evaluates the potential influence of the Montreal Formula 1 Grand Prix on high-speed collision rates in Quebec from 2000 to 2022. We explore whether temporal spikes in dangerous driving behavior are associated with the F1 event, using a variety of causal inference and time-series modeling approaches. Principal investigator behind the project: Dr. Ignacio Nazif-Muñoz.
Project Structure
Weather Variables and Event Setup
- Identifies Grand Prix dates by year and determines availability based on whether the event occurred.
- Defines pre/post-event time windows and day-of-week controls.
- Links these periods to weather data from nearby meteorological stations.
- Merges with high-speed collision data from provincial sources.
Time-Series Construction and Exploration
- Collapses weather and collision variables into structured time-series datasets.
- Evaluates temporal dynamics including autocorrelation and seasonality.
- Prepares inputs for advanced modeling in later steps.
Statistical Modeling: Case-Crossover and DiD
- Implements a stratified case-crossover design to control for time-invariant confounding.
- Uses Synthetic Control Methods (SCM) to build counterfactual trajectories for high-speed collisions.
Causal Inference Using Google’s CausalImpact package
- Estimates the causal impact using a Bayesian time series analysis of the F1 event across different years and windows.
Post review analyses and additions
- Post review adjustments and additional analyses based on peer feedback.