This project combines Data Science and Operations Research to improve driver allocation in a ride-hailing system. The goal is to predict ride demand and allocate drivers efficiently based on those predictions.
Ride demand varies throughout the day, while the number of available drivers is limited. This creates a need to allocate drivers in a way that maximizes coverage and reduces inefficiency.
- Built a simple Linear Regression model using hour as a predictor
- Built an improved Random Forest model using additional features such as pickup location
- Model comparison:
- Simple Model: R² ≈ 0.44
- Improved Model: R² ≈ 0.91
- Total drivers available: 500
- Each driver works 8 hours per day
- Driver-hours are distributed proportionally based on predicted demand
This ensures:
- More drivers during peak hours
- Fewer drivers during low-demand periods
- Improved demand prediction significantly enhanced allocation decisions
- Driver allocation follows demand patterns
- More efficient use of limited driver resources
data/ - input dataset
notebooks/ - data analysis and modeling
results/ - predictions and allocation outputs
Accurate demand prediction enables better operational decision-making by improving how limited resources are allocated across time.