🚖 Uber Booking Analysis

📌 Project Overview

This project focuses on analyzing Uber ride booking data using Data Science and Machine Learning techniques. The goal is to extract insights from ride data and build predictive models to determine whether a ride will be completed or not.

---

🎯 Problem Statement

Uber booking systems generate large volumes of data related to rides, customers, and operations. Analyzing this data helps improve:

• Ride completion rates
• Customer satisfaction
• Operational efficiency

This project aims to:

• Clean and preprocess the dataset
• Perform Exploratory Data Analysis (EDA)
• Handle class imbalance
• Build and evaluate machine learning models

---

📊 Dataset Description

The dataset contains ride-related attributes such as:

Column Name	Description
Booking ID	Unique ride identifier
Booking Value	Total fare of ride
Ride Distance	Distance traveled
Driver Ratings	Rating given to driver
Customer Rating	Rating given by customer
Payment Method	Cash / UPI / Card
Vehicle Type	Type of vehicle
Booking Status	Completed / Cancelled

---

🛠️ Technologies Used

• Python 3
• Pandas
• NumPy
• Matplotlib
• Seaborn
• Scikit-learn
• PySpark

---

🔧 Data Preprocessing

The dataset was cleaned using the following steps:

• Handling Missing Values

  • Numerical → Mean
  • Categorical → Mode

• Removing Duplicates

• Outlier Detection

  • IQR Method
  • Z-score Method

• Feature Scaling

  • Min-Max Scaling
  • Standardization

---

📈 Exploratory Data Analysis (EDA)

Techniques used:

• Histogram
• Scatter Plot
• Boxplot
• Heatmap

🔍 Key Insights

• Most bookings are Completed
• Booking value increases with ride distance
• Data shows slight skewness
• Ratings influence ride completion

---

⚖️ Handling Class Imbalance

Problem:

• Completed bookings >> Non-completed bookings

Solution:

• SMOTE (Synthetic Minority Oversampling Technique)
• Balanced dataset improved model performance

---

🤖 Machine Learning Models

The following models were implemented:

1. Logistic Regression

• Used for binary classification
• Predicts probability of ride completion

2. K-Nearest Neighbors (KNN)

• Based on similarity between data points

3. Naive Bayes

• Probabilistic classifier
• Fast and efficient

---

📊 Model Evaluation

• Train-Test Split (70-30)
• Cross Validation
• Confusion Matrix
• Accuracy Score
• Z-Test for validation

---

🏆 Results

• Multiple models were compared

• Best Model: Decision Tree Classifier

  • Highest accuracy
  • Captures complex patterns

⚠️ Limitation

• May overfit on data

✅ Alternative

• Naive Bayes provides more stable performance

---

📌 Conclusion

This project demonstrates how data science techniques can be applied to real-world ride booking systems. The analysis helps in:

• Improving ride completion prediction
• Understanding customer behavior
• Enhancing operational decisions

---

📂 Project Structure

Uber-Booking-Analysis/
│── UberBookingAnalysis.ipynb
│── README.md
│── dataset.csv (optional)

---

🚀 How to Run

1. Clone the repository
2. Open the notebook in Google Colab or Jupyter
3. Run all cells

---

👨‍💻 Authors

• Roll No: 13, 14, 15

---

📅 Date

April 2026