/guzmanwolfrank.github.io

webportfolio site

Primary LanguageHTMLMIT LicenseMIT

UK Traffic Accident Analysis (2015-2018)

UK Traffic Study Banner

Overview

A comprehensive analysis of UK road safety data from 2015-2018, examining over 529,294 traffic accidents to understand patterns, risk factors, and potential intervention points for improving road safety.

License: MIT Python Tableau

Table of Contents

Data Sources

The analysis utilizes three main datasets:

  • Accidents Data (2015-2018)
  • Casualties Data (2015-2018)
  • Vehicles Data (2015-2018)

Total records analyzed: 529,294 accidents

  • Fatal Accidents: 6,658
  • Serious Accidents: 87,462
  • Slight Accidents: 435,174

ETL Process

  1. Data Validation and Structure Assessment

    • Performed initial data comparison between redundant datasets
    • Validated matching data between directories
    • Confirmed 100% match for all datasets

  2. File Structure Standardization

    • Implemented consistent naming conventions
    • Standardized file organization
    • Created unified directory structure

  3. Data Transformation

    • Converted timestamps from UK to US format
    • Resolved time format inconsistencies
    • Implemented datetime validation checks

  4. Data Cleaning

    • Executed custom cleaning scripts
    • Performed error correction
    • Ensured format consistency
    • Generated comprehensive cleaning reports

  5. Data Consolidation

    • Created master files for:
      • Accidents
      • Casualties
      • Vehicles

Analysis & Findings

1. Temporal Analysis

  • Peak accident times during rush hours (7-9 AM and 4-6 PM)
  • Higher severity rates during nighttime (11 PM - 4 AM)
  • Distinct weekend vs weekday patterns

2. Weather Impact

  • Adverse weather significantly affects accident severity
  • Rain: Higher frequency, lower average severity
  • Snow/Ice: Lower frequency, higher severity rates

3. Road and Speed Analysis

  • Single carriageways show highest accident rates
  • Strong correlation between speed limits and severity
  • Urban roads: High frequency, lower severity
  • Motorways: Low accident rates despite high speeds

4. Casualty Analysis

  • Young adults (18-25): Higher representation
  • Elderly (65+): Higher severity rates
  • Distinct pedestrian and cyclist patterns

Visualizations

Hourly Distribution Monthly Severity Trends Severity Heatmap Time of Day Analysis

Interactive Dashboards

Access the interactive Tableau dashboards:

Technologies Used

  • Python 3.8+
  • Pandas
  • NumPy
  • Matplotlib
  • Seaborn
  • Tableau
  • Jupyter Notebook

Installation

# Clone the repository
git clone https://github.com/guzmanwolfrank/uk-traffic-analysis.git

# Navigate to project directory
cd uk-traffic-analysis

# Install required packages
pip install -r requirements.txt

Usage

# Example code for loading the datasets
import pandas as pd

# Load the accident data
accidents_df = pd.read_csv('data/accidents_master.csv')

# Load the casualties data
casualties_df = pd.read_csv('data/casualties_master.csv')

# Load the vehicles data
vehicles_df = pd.read_csv('data/vehicles_master.csv')

Recommendations

  1. Smart Infrastructure Implementation (15-20% potential reduction)

    • AI-powered traffic management
    • Dynamic speed limits
    • Connected vehicle infrastructure

  2. Enhanced Education Programs (10-15% potential reduction)

    • Continuous learning systems
    • Virtual reality hazard training
    • Vulnerable user awareness

  3. Technology-Based Solutions (20-25% potential reduction)

    • Advanced driver assistance systems
    • Vehicle-to-vehicle communication
    • Automated emergency braking

License

This project is licensed under the MIT License - see the LICENSE.md file for details.

Contact

Wolfrank Guzman

Acknowledgments

  • UK Department for Transport for providing the accident data
  • Contributors and reviewers who helped improve this analysis
  • The open-source community for the tools and libraries used