# Hotel Booking Cancellation Analysis

## Overview
This project analyzes hotel booking data to better understand what drives cancellations.

The goal was to explore patterns in customer behavior and identify factors that make a booking more or less likely to be canceled. Rather than focusing only on description, the analysis aims to provide insights that could be useful from a business perspective.

---

## Dataset
The dataset is based on the "Hotel Booking Demand" dataset.

It includes over 100,000 booking records with features such as:
- Lead time (how early the booking was made)
- Hotel type (city or resort)
- Customer type (repeated or new)
- Average daily rate (ADR)
- Cancellation status

The dataset is large enough to reveal meaningful patterns, while still being manageable for exploratory analysis.

---

## Data Preparation & Cleaning

Before starting the analysis, the dataset was reviewed and cleaned:

- Duplicate records were identified and removed to avoid bias in the results.
- Missing values were examined carefully. In some columns (such as company or agent), missing values represent real situations rather than errors, so they were not removed.
- Several variables contain many zero values (for example, children or previous cancellations). These were kept, as zero is a meaningful value and reflects actual behavior.
- Outliers were identified in variables such as lead time and ADR. These were not removed, as they may represent real extreme cases. Instead, they were handled carefully in the visualizations to keep the plots readable.

The cleaning approach focused on preserving the original structure of the data while improving consistency.

---

## Descriptive Statistics

Descriptive statistics were used to get an initial understanding of the data.

There is noticeable variation across key features such as lead time and ADR, which suggests that booking behavior differs significantly between customers.

---

## Analysis & Insights

### 1. Cancellation Rate by Hotel Type


City hotels show a higher cancellation rate compared to resort hotels.

This may indicate that bookings in city hotels are less stable, possibly due to shorter trips, business travel, or more flexible planning.

---

### 2. Lead Time and Cancellation


Canceled bookings tend to have a higher average lead time.

This suggests that reservations made far in advance are less certain, and customers are more likely to change their plans.

---

### Additional Insight: Lead Time Ranges


Looking at lead time in ranges provides a clearer picture.

Cancellation rates increase gradually as lead time increases, which confirms that the pattern is consistent and not driven only by extreme values.

---

### 3. Repeated vs First-Time Guests


Repeated guests have a noticeably lower cancellation rate compared to first-time guests.

This suggests that returning customers are more committed and have higher confidence in their bookings.

---

### 4. Price (ADR) and Cancellation


Higher-priced bookings show greater variability and are slightly more likely to be canceled.

One possible explanation is that customers making more expensive bookings are more sensitive to changes and may continue comparing alternatives after booking.

---

## Business Insights

Based on the analysis, several practical implications can be suggested:

- Bookings made far in advance should be monitored more carefully, as they carry a higher risk of cancellation.
- City hotels may benefit from stricter cancellation policies or incentives that increase commitment.
- Returning customers appear more reliable, so investing in loyalty programs could reduce cancellations.
- Higher-priced bookings may require additional confirmation steps or incentives to reduce uncertainty.

---

## Conclusion

The analysis shows that cancellations are not random.

They are influenced by timing, pricing, and customer experience. Understanding these patterns can help businesses improve forecasting, reduce cancellations, and make better operational decisions.

---

## Data Source

The dataset used in this project is based on the **"Hotel Booking Demand" dataset**.

Source:
https://www.kaggle.com/datasets/jessemostipak/hotel-booking-demand

The data was used for academic purposes, and further cleaned and analyzed as part of this project.