Hotel Booking Cancellation Prediction IDS f24 – Introduction to Data Science Semester Project

Project Overview

Hotel booking cancellations are a major challenge for the hospitality industry, leading to revenue loss and inefficient resource planning. This project analyzes hotel booking data and predicts whether a reservation will be canceled or checked-in using machine learning.

The project demonstrates the complete data science lifecycle:

• Data cleaning and preprocessing
• Exploratory Data Analysis (EDA)
• Machine Learning model training
• Interactive deployment using Streamlit

Objectives

• Identify key factors influencing booking cancellations
• Perform detailed exploratory data analysis
• Train and evaluate a Random Forest classifier
• Build an interactive web application for real-time predictions

Project Structure

hotel-booking-project/ ├── data/ │ └── hotel_bookings.csv ├── models/ │ └── hotel_cancellation_model.pkl ├── src/ │ └── app.py ├── requirements.txt └── README.txt

Dataset

• Approximately 119,000 hotel bookings
• Target variable: is_canceled
• Features include lead time, ADR, market segment, deposit type, country, previous cancellations, and special requests

Machine Learning Model

• Algorithm: Random Forest Classifier
• Handles non-linear relationships and mixed data types
• Outputs probability of booking cancellation

Streamlit Application

The app includes:

• Home page with dataset overview
• Interactive EDA visualizations
• Prediction form with clean UI controls
• Highlighted prediction results with probabilities
• Key takeaways section summarizing insights

Installation & Usage

1. Install dependencies: pip install -r requirements.txt

2. Run the app: streamlit run src/app.py

Key Takeaways

• Lead time strongly affects cancellation likelihood
• Previous cancellations increase risk
• Deposit type and market segment matter
• Random Forest performs well for this task

Course Information

Course: Introduction to Data Science (IDS f24) Project Type: Semester Project