2025-12-20

Complexity of Road Safety Modeling

Evaluating the 1983 UK Seatbelt Law isn’t simple. Raw data is often clouded by:

  • Economic Shifts: Changes in petrol prices influence how much people drive.
  • Policy Impact and Shortcomings: Law had an impact on front-seat safety, but might have led to different behaviors that affected rear-seat passengers differently.
  • Seasonality: Winter vs. Summer driving patterns.
  • Exposure: Total distance driven (kms).
  • The Paradox: How to separate long-term safety improvements from the actual impact of the law, or economic factors.




Strategic Design: The Dashboard Solution

Our Seatbelts_Dashboard App provides a modular analytical pipeline to address specific challenges in road safety modeling:


Tab Name Analytical Problem Solution Provided
Exploratory Unclear Correlations Real-time univariate regression mapping.
Front vs Rear Attribution Bias Comparison of legislated vs. non-legislated groups.
Seasonality Environmental Noise Monthly factor controls and Interaction Toggles.
Prediction Static Reporting Live “What-If” simulator with multivariant model and Time-Trend correction.



App : https://rpubs.com/MAlShawa/1382929.

Presentation and App Code: https://github.com/MAlShawa/Seatbelts_Dashboard_App.

Usability & Statistical Integrity

The Seatbelts_Dashboard App is built for both stakeholders and statisticians, balancing ease of use with deep mathematical rigor.

  • Analytical Accuracy: Provides model details, and resolves the “Distance Paradox” by accounting for long-term safety trends via a Time-Trend variable.
  • Statistical Safeguards: Built-in Durbin-Watson tests to check for autocorrelation, while Residual Plots validate model assumptions.
  • Strategic Value: Empowers users to predict fatalities by manipulating petrol prices, distance, and law status.
  • Modern UI Design: Utilizing the lumen professional theme, high-contrast bold labels, and interactive UI elements such as sliders and checkboxes.
  • Clarity: Uses numbered seasonal legends to bridge the gap between complex plots and the prediction simulator.
  • Data Explained: Features a live “About Data” section with full variable description.
  • Ease of Use: Features a live “About App” section with a Navigation Guide and description of the UI Components.

Conclusion: Data-Driven Decisions

This dashboard transforms the historical Seatbelts dataset into a dynamic modern decision-support tool for policy evaluation.

  1. Exploratory: Visualizing raw trends, the affect of seasonality, and the effect of policy on legislated vs. non-legislated groups.
  2. Validating: Using Durbin-Watson tests for autocorrelation, and Residual Plots to validate model assumptions.
  3. Predictive: Simulating the future of road safety by allowing the decision maker to interact with the model and see the effect of changing its predictors’ values.

The application provides:

  1. User-Centric UI: Easy to use controls, such as bold toggles for the Law, sliders for Seasonality, and precise Kilometer inputs.
  2. Statistical Honesty: The App doesn’t just provide a number; it provides the Mean Absolute Error (MAE) and Adjusted R².
  3. Visual Feedback: The interface uses a clean, professional theme to make complex econometrics accessible to stakeholders.
Drive safer policy through interactive econometrics.