Hotel Booking Analytics & Cancellation Prediction

Project Overview

This project analyzes hotel booking behavior and builds a machine learning model to predict booking cancellations.

The project combines:

• Business analytics using an Excel dashboard
• Machine learning models using Orange Data Mining

The goal is to understand booking trends, identify factors affecting cancellations, and evaluate predictive models.

Key Dashboard Insights

• Peak booking demand occurs during August
• Cancellation rates exceed 40% during April–June
• Group bookings show the highest cancellation rate
• Direct and Online TA channels generate the highest ADR

Machine Learning Model (Orange Data Mining)

A machine learning workflow was built using Orange Data Mining to predict whether a booking will be cancelled.

Models tested:

• Logistic Regression

• Random Forest

  

Model Performance

Random Forest performed better than Logistic Regression across most evaluation metrics.

Model	AUC	Accuracy	F1 Score	Precision	Recall
Random Forest	0.922	0.862	0.860	0.862	0.862
Logistic Regression	0.863	0.811	0.804	0.814	0.811

Confusion Matrix

ROC Curve

Project Structure

hotel-booking-analytics
│
├── data
│   └── hotel_booking.csv
│
├── dashboard
│   └── README.md
│
├── orange-model
│   ├── hotel_booking_prediction.ows
│   └── README.md
│
├── screenshots
│   ├── dashboard.png
│   ├── orange_workflow.png
│   ├── confusion_matrix.png
│   └── roc_curve.png
│
└── README.md