Airline Booking Demand Forecasting

Objective

The goal of this project is to develop a simple, interpretable forecast model for airline bookings that minimizes the Mean Absolute Scaled Error (MASE) compared to a naïve baseline. We focus on analytics and business logic, not machine learning, to keep the approach transparent and practical.

Data Description

We use two datasets:

• Training (sample): Cumulative bookings for various departure dates, with each departure having a record of 60 prior days' bookings.
• Validation: Similar structure, used to test our forecasting logic.

Approach

1. Exploratory Data Analysis (EDA)

• We start by checking for missing values and understanding the distribution of bookings.
• Visualizations show how bookings vary by day of week and by days prior to departure.

2. Feature Engineering

• We extract features like day_of_week and DaysToDepart (as a numeric value).
• For each departure, we calculate the final demand (bookings on the day of departure), remaining demand, and booking rate.

3. Forecasting Models

• Additive Model: Forecast = cumulative bookings + mean remaining demand (from training data).
• Multiplicative Model: Forecast = cumulative bookings / mean booking rate (from training data).
• Both models are run in two ways: by days prior only, and by days prior + day of week.
• Weighted Average: We also try a weighted average of the two models, tuning the weight to minimize MASE.

4. Evaluation

• We use MASE as our main metric, comparing each model's forecast to the naïve forecast provided in the validation set.
• The notebook prints out MASE scores for all approaches and highlights the best result.

5. Visualization

• The notebook includes plots for EDA and for comparing actual vs. forecasted demand for a sample departure date.

How to Run

1. Make sure you have Python 3 installed.
2. Install all dependencies with:

   pip install -r requirements.txt
   

3. Place airline_data_training.csv and airline_data_validation.csv in the project directory.
4. Open ForecastFlightBookings.ipynb in Jupyter Notebook or VS Code.
5. Run the notebook cells in order. You'll see EDA, model results, and visualizations.

Notes

• This project is intentionally simple and transparent—no machine learning, just analytics and business logic.
• The code is well-commented and organized for clarity.
• If you want to extend the analysis, you can add more features, try other error metrics, or experiment with different weighting schemes.