Eula_Portfolio

Motorcycle Accident Risk Analysis (California)

Skills: Data Cleaning

Exploratory Data Analysis

Data Visualization

Large Dataset Processing

SQL/SQLite

Statistical Analysis

R (dplyr, ggplot2)

Overview

Motorcycle accidents account for a significant portion of traffic injuries and fatalities each year. This project analyzes large-scale traffic collision data to identify demographic, environmental, and temporal patterns associated with motorcycle accidents. The goal is to uncover patterns that may inform safety interventions and policy improvements.

Motivation

As a night-shift medical laboratory scientist, I frequently encounter trauma cases resulting from motorcycle accidents. This exposure motivated an investigation into whether large-scale traffic datasets could reveal patterns associated with higher accident risk or severity.

Datasets

This analysis integrates three public transportation datasets:

1. California Traffic Collision Data (SWITRS)

• 9.4 million collision records
• 18.7 million involved parties
• Variables include vehicle type, road conditions, lighting, injury severity, and rider demographics.

2. Road Traffic Injury Dataset

• 449,850 observations
• Contains injury and fatality statistics across transportation modes.

3. U.S. Transportation Accident Data

• National crash statistics from 1988–2022.

Data Preparation

Data preparation involved several steps:

• Importing large datasets including SQLite traffic collision data
• Filtering relevant motorcycle-related records
• Joining collision and party tables
• Replacing missing values with standardized labels
• Removing duplicate observations
• Preparing variables for exploratory analysis

Handling large datasets required memory management and selective filtering to focus on motorcycle-related incidents.

Analysis Methods

The analysis focused on exploratory data analysis and statistical comparisons, including:

• Age distribution analysis
• Gender comparisons
• Environmental condition analysis (lighting, weather, road conditions)
• Mode-of-transport comparisons
• Injury and fatality rate calculations

Visualization techniques included:

• Bar charts
• Histograms
• Boxplots
• Comparative mode-of-transport charts

Key Findings

Fatality Risk

Motorcycle accidents resulted in fatalities approximately 3.2% of the time, with injuries occurring in 82% of cases.

Age Patterns

Younger riders, particularly those in their twenties and early thirties, were involved in the highest number of accidents.

Gender Differences

Motorcycle accidents were heavily skewed toward male riders, though some records contained unknown gender classifications.

Environmental Conditions

Contrary to common assumptions, most motorcycle accidents occurred under normal weather, road, and lighting conditions, suggesting that rider behavior and traffic interactions may be more influential than environmental conditions.

Comparison to Other Modes

Motorcycle accidents produced a higher injury rate than trucks, buses, and bicycles, but slightly lower than cars and pedestrians in overall injury distribution.

Implications

The findings highlight potential areas for safety intervention:

• Targeted rider safety programs for younger riders
• Improved rider training and licensing requirements
• Enhanced data collection for rider demographics
• Infrastructure and safety initiatives focused on motorcycle riders

Tools

R, ggplot2, dplyr, tidyr, RSQLite, readxl