Uber Traffic Analysis & Prediction - Internship project

Project Overview

This project analyzes historical traffic data and builds predictive models to forecast hourly traffic volume at road junctions. The objective is to support ride-hailing platforms like Uber in optimizing routes, reducing congestion impact, and improving ETA accuracy.

---

Objective

• Analyze traffic patterns across time and junctions
• Engineer time-based features
• Build and evaluate machine learning and deep learning models for traffic forecasting

---

Approach

• Data Cleaning & Feature Engineering

• Exploratory Data Analysis (EDA)

• Modeling

  • Linear Regression
  • Random Forest
  • Gradient Boosting
  • LSTM (Deep Learning)

• Evaluation Metrics**

  • MAE, MSE, R²

---

Tech Stack

• Python
• Pandas, NumPy
• Matplotlib, Seaborn
• Scikit-learn
• TensorFlow / Keras

---

Key Insights

• Strong hourly and daily traffic patterns identified
• Ensemble and LSTM models outperform baseline regression
• Framework scalable for real-world traffic forecasting systems

---

Repository Structure

Uber-predictive-traffic-modelling/
│
├── Dataset_Uber_Traffic.xlsx
├── hyderabad_weather_2015_2017.xlsx
└── hyderabad_all_special_events_2015_2017_merged.xlsx
│
├── Uber_analysis_final.ipynb
│
└── README.md

---

Future Work

• Real-time data integration
• Model deployment (Flask / FastAPI)
• Live traffic dashboard

---

Author

Mohan Chand Kommalapati

Aspiring Data Scientist | Predictive Analytics Enthusiast