🚕 Uber Ride Requests Analysis and Clustering

• This project analyzes Uber ride request data to uncover demand patterns and driver behavior using Python.
• It includes cleaning, feature engineering, visualizations, and clustering of pickup locations with K-Means.

---

📚 Table of Contents

• 📌 Description
• 🔄 Project Workflow
• 💡 Key Insights
• 🛠️ Recommendations
• 🗂️ Dataset
• ⚙️ Installation
• 🚀 Run the Full Pipeline
• 📽️ Inference Demo
• 📁 Project Structure
• 👥 Contributors
• 📝 License

---

📌 Description

A comprehensive end-to-end data science project analyzing real-world Uber ride request data using Python.

---

🔄 Project Workflow:

• Data Cleaning:

  • Parsed and standardized datetime columns (Request timestamp, Drop timestamp).
  • Handled missing values in Driver id and Drop timestamp without dropping large portions of data.
  • Cleaned categorical fields like Status and Pickup point (e.g., stripping whitespace, title casing).

• Feature Engineering:

  • Extracted informative time-based features:
    • Request hour, Request day, and Request Date
    • Trip duration (as timedelta and in minutes)
    • Time slot classification (Morning, Afternoon, Evening, Night)
    • Boolean flags like Driver Available and Is Completed

• Data Visualization:

  • Identified:
    • Demand peaks (morning and evening)
    • Driver inactivity periods
    • Request trends by hour, day, and location
    • Cancellation patterns and completion rates

• Clustering with K-Means (k=5):

  • Simulated pickup coordinates and applied clustering to segment NYC into 5 operational zones.
  • Visualized high-density clusters vs. low-demand zones.

---

💡 Key Insights:

• Peak Demand: Evening hours (5–9 PM) have highest request volume but lowest driver availability.
• Driver Inactivity: Fridays show high inactivity despite high demand.
• Trip Duration: Early morning trips (12–4 AM) are slightly longer.
• Top Drivers: Most top-10 drivers completed between 20–22 trips — tight distribution.
• Pickup Patterns: Majority of requests originate from the City; Airport has high rate of "No Cars Available".

---

🛠️ Recommendations:

• Incentivize drivers during evening peaks and on inactive days (e.g., Friday).
• Promote early morning hours (12–3 AM) with offers and safety campaigns.
• Use K-Means zones to operationally split NYC for optimized driver distribution.
• Coordinate with airports for better pickup logistics & reduce city cancellations.

---

🗂️ Dataset

• File Name: Uber Request Data.csv
• Source: Provided by course instructors
• Format: Tab-separated values (.tsv)
• Rows: 6,745 Uber ride requests
• Description: Raw log of Uber rides captured over a period of several days. Used for analyzing ride demand, driver activity, cancellations, and trip patterns.

🔑 Key Columns:

• Request id: Unique numeric ID for each ride request
• Pickup point: Location where the ride was requested from (City or Airport)
• Driver id: ID of the driver assigned (may be missing if unassigned)
• Status: Trip outcome - Trip Completed, Cancelled, or No Cars Available
• Request timestamp: Timestamp when the user requested the ride
• Drop timestamp: Timestamp when the trip was completed (may be null)

---

⚙️ Installation

Requirements

To get started, make sure you have Python 3.10+ installed on your system.
Then install all required libraries by running:

pip install -r requirements.txt

---

📦 Core Libraries Used

• pandas → data manipulation
• numpy → numerical operations
• matplotlib, seaborn → visualizations
• scikit-learn → KMeans clustering
• python-dateutil → parsing flexible timestamps

---

✅ Setup Checklist

Ensure the following are in place before running the project:

• data/Uber Request Data.csv is inside the data/ folder
• src/ contains all modular code files
• output/ folder exists for saving processed results
• main.py exists in the project root

---

🚀 Run the Full Pipeline

Run the entire analysis and visualization pipeline with:

python main.py

---

📽️ Inference Demo

Once the project is launched via main.py, the following steps are performed:

• Loads and parses the raw Uber request dataset
• Cleans and preprocesses timestamps, missing values, and categories
• Extracts useful features like time slot, duration, and weekday
• Adds simulated geo-coordinates and clusters pickup points using K-Means
• Displays visual insights (trip patterns, demand peaks, hotspot regions)
• Saves enriched dataset to output/Uber with features.csv

Visualizations appear live using matplotlib and seaborn
Final processed dataset is reusable for further analysis

---

📁 Project Structure

The project is organized in a modular, production-ready structure:

Uber-Trip-Analysis/
│
├── data/                         # Raw input data (Uber Request Data.csv)
│   └── Uber Request Data.csv
│
├── output/                       # Final enriched data & visual assets
│   └── Uber with features.csv
│
├── src/                          # All modular processing code
│   ├── __init__.py
│   ├── data_loading_and_exploration.py
│   ├── data_cleaning_and_preprocessing.py
│   ├── data_feature_engineering.py
│   ├── data_visualization.py
│   ├── location_clustering.py
│   └── save_transformed_data.py
│
├── main.py                       # Master pipeline to run the entire workflow
├── requirements.txt              # Python dependencies
└── README.md                     # Project documentation

---

👥 Contributors

• Malak Salem: Data loading, exploration, and reporting key patterns.
• Laila Shawky: Cleaning and feature engineering (timestamps, nulls, derived fields).
• Jumanah Rushdi: Visualizations and spatial clustering (K-Means on pickup points).

---

📝 License

• Open source under the MIT License.
• Free to use, modify, and distribute with attribution.