-British-Airways-Predictive-Model

This repository contains a predictive modeling project for British Airways to forecast customer bookings using historical booking data.

Project Description

The objective of this project is to build a machine learning model to predict the likelihood of a customer completing a booking. The model uses various features such as purchase_lead, flight_hour, length_of_stay, and more to make predictions. The results are used to enhance customer engagement and operational efficiency.

Dataset

The dataset used for this project is customer_booking.csv. It contains various features related to customer bookings.

Files

• customer_booking.csv: The dataset used for the project.
• notebook.ipynb: Jupyter notebook with the data exploration, preprocessing, model training, and evaluation steps.
• feature_importance.png: Visualization of the top 20 feature importances.
• presentation.pptx: PowerPoint presentation summarizing the findings of the project.

How to Run

1. Clone this repository:

   git clone https://github.com/YOUR_GITHUB_USERNAME/YOUR_REPOSITORY_NAME.git

2. Navigate to the project directory:

   cd YOUR_REPOSITORY_NAME

3. Install the required packages:

   pip install -r requirements.txt

4. Open the Jupyter notebook:

   jupyter notebook notebook.ipynb

Results

• Accuracy: 85%
• ROC AUC Score: 0.78

Key Features

• purchase_lead: The number of days in advance the booking was made.
• flight_hour: The hour of the flight.
• length_of_stay: Duration of stay.
• flight_day: The day of the week when the flight is scheduled.
• num_passengers: Number of passengers in the booking.

License

This project is licensed under the MIT License - see the LICENSE file for details.