Food Demand Forecasting Dataset

Data Provided:

• Train Set:
  • Weekly Demand Data (train.csv): Contains historical demand information from Week 1 to Week 145 for various center-meal combinations.
  • fulfilment_center_info.csv: Detailed information on fulfillment centers, including location and capacity.
  • meal_info.csv: Features of each meal such as category, subcategory, price, and any applicable discounts.
• Test Set:
  • test.csv: Includes all the necessary features, excluding the demand values, for predictions.

Evaluation Metric:
Performance will be measured using the Root Mean Squared Logarithmic Error (RMSLE), which focuses on the relative error between predicted and actual demand. The final evaluation is based on minimizing this error across all test cases.

Key Insights for Modeling:

1. Perishability: The perishability of raw materials makes precise demand forecasting more challenging, and small prediction errors can result in financial losses.
2. City-Specific Factors: Each city and fulfillment center has unique attributes that may affect demand, including local preferences, weather patterns, or regional events.
3. Discount Effects: The model needs to account for pricing and discount strategies that influence customer behavior and, consequently, demand patterns.

Additional Considerations:

• Seasonality: Consider the seasonal variations in food demand, which could significantly influence purchasing behavior.
• Customer Trends: Changes in customer preferences, health trends, or new meal offerings may affect forecasts.
• External Factors: Holidays, events, and macroeconomic conditions could also shift demand and should be factored into the model.