New Food Delivery Time Prediction Model

Requirements Clarification & Assessment

1. Understanding the Model's Scope:

   • Delivery Components: Does the model consider all components of delivery time, such as finding a rider, kitchen preparation, and traffic conditions? Are these components sequential?
   • Historical Data Availability: Is there access to historical delivery data for testing and validation?
   • Real-time Updates: Does the model update predictions in real-time as new information becomes available (e.g., when a rider picks up the order)?
   • Deployment Stage: Is the model still in development, or has it been deployed?

2. Objective of the Model:

   • Accuracy Goal: Is the primary goal to improve the accuracy of estimated delivery times (EDT)?
   • Replacement: Will the new model replace the old one if it proves more accurate?

3. Data Logging:

   • Activity Logging: Can all activities, such as time to find a rider, order readiness, and pickup/delivery times, be logged for analysis?

4. Comparison Metrics:

   • Performance Metrics: What metrics will define success (e.g., RMSE, MAE, MSE)?
   • Segment Evaluation: Are there specific segments (e.g., cuisine type, geography) where performance evaluation is critical?