Booking Prediction Model

Overview

This project aims to predict customer booking behaviors by classifying them into three categories:

1. Booked and Canceled
2. Booked and Checked Out
3. Booked and Did Not Show

By analyzing these categories, we can identify potential loyal customers and tailor targeted marketing strategies accordingly.

Table of Contents

• DataWareHouse_Design
• ETL_Process
• Azure
• Data Preparation
• Feature Selection
• Model Selection
• Training and Scoring
• Evaluation Metrics
• Power BI
• Usage
• Conclusion
• License

DataWareHouse_Design

This file contains the data warehouse schema design for a hotel booking system. The schema follows a star schema model, with fact tables and dimension tables that support efficient queries for booking, property operations, and revenue analysis. Read More

ETL_Process

This project implements an ETL (Extract, Transform, Load) process that extracts data from various file sources, transforms it to fit the desired format, and loads it into a SQL Server database. Read More

Azure

This project leverages Azure Cloud Services for managing and processing data, integrating data flows into Azure Machine Learning for model building and analytics. Read More

Data Preparation

In this stage, we focused on ensuring the quality of our data:

• Missing values in the rating column were replaced using the median to maintain the integrity of the dataset.

Feature Selection

We selected specific features to build the model:

• Features Used:
  • property_id
  • check_in_date
  • checkout_date
  • booking_date
  • no_guests
  • room_category
  • booking_platform
  • ratings_given
  • booking_status
  • revenue_generated
  • revenue_realized

• The dataset was split into a training set (70%) and a test set (30%) to validate the model's performance.

Model Selection

We utilized Multiclass Logistic Regression to classify customer behaviors into the three specified categories. This model was chosen for its effectiveness in handling multiclass classification tasks.

Training and Scoring

• The model was trained using the training dataset to learn patterns and relationships within the data.
• The remaining 30% of the data was used for scoring to evaluate the model's predictive accuracy.

Evaluation Metrics

To assess the model's performance, we utilized the following metrics:

• Overall Accuracy: Measures the proportion of correct predictions.
• Micro Precision: Averages precision across all instances, treating them equally.
• Macro Precision: Averages precision for each class independently, providing insights into performance on minority classes.

Power_Bi

This project leverages Power BI to visualize and analyze hotel booking data. The dashboard provides key performance indicators (KPIs), enabling stakeholders to monitor and optimize hotel performance. Read More

Usage

To use the model:

• Clone this repository.

Conclusion

This project highlights the potential for predicting customer behaviors in the booking industry. By understanding customer patterns, businesses can enhance their marketing strategies and improve customer loyalty.