R for Transportation Data Analytics

Pennsylvania Road Lighting and Safety

This repository demonstrates an end-to-end analysis of the question:

How does roadway lighting affect roadway safety?

This project demonstrates the use of the R statistical programming languaeg and several packages to conduct a "data science" analysis of the relationship between roadway lighting (i.e. streetlights) and crashes on major roads in Pennsylvania.

View the R notebook here: https://markegge.github.io/safety-demo/

Concepts Demonstrated

This package illustrates several common transportation data analysis tasks, including:

• Managing data using the fast data.table package
• Using the sf package to perform a spatial join between point data (crashes) and line data (road segments)
• Using the rpart package to predict lighting conditions for road segments with unknown lighting conditions based on road segments with known lighting conditions
• Exploratory data analysis using the leaflet mapping package and ggplot2 plotting package
• Fitting a zero-inflated Poisson regression model to predict crash counts controlling for traffic volume

Dataset

This project combines crash data with road inventory and traffic data:

• Road segment and traffic count data from PennDOT's RMS Traffic Open Data
• PA Crash Data from PennDOT's Public Crash Databases

Attribution

This project is created as a demonstration for CCAC ATE-252 and may be freely reused and distribuetd.

Project author: Mark Egge (mark@eateggs.com)