Allegheny County Right-Turn & RTOR Crash Analysis (2005–2024)

This repository analyzes 20 years of PennDOT crash data (2005–2024) for Allegheny County, PA, with a specific focus on:

• Right-turn crashes
• Right-Turn-On-Red (RTOR) crashes
• Non-motorist involvement (pedestrians & cyclists)
• Crash severity and fatalities
• County-wide spatial risk mapping and heatmaps

This project supports Vision Zero–style safety analysis by identifying high-risk intersections and movement patterns that disproportionately impact vulnerable road users.

---

Data Sources & Coding Standards

• Data Source: PennDOT Crash Reporting System

• Tables Used:

  • Crash Table: crash-level severity, location, non-motorist counts
  • Vehicle Table: per-vehicle movement codes
  • Roadway Table: roadway & street name context

• Key Movement Codes:

  • 09 = Right Turn on Red (RTOR)
  • 10 = Right Turn
  • 11, 12 = Left Turn

• Intersection Classification:

  • INTERSECT_TYPE = 01–13 → intersection-related
  • 00 → non-intersection
  • 99 → unknown (excluded)

---

Automatic Multi-Year Data Loading

• Automatically scans:
  • Allegheny_2005 through Allegheny_2024
• Automatically loads:
  • Crash records
  • Vehicle movement records
  • Roadway & street name data
• Automatically merges:
  • Crash + Vehicle + Roadway tables
• No manual year selection is required

---

Non-Motorist Risk Analysis

• Left-turn vs Right-turn vs RTOR crash comparison
• Non-motorist hit rate calculation
• Pedestrian vs cyclist breakdown
• Serious injury and fatality rate computation
• Motorist vs non-motorist severity comparison

Confirmed Findings

• Right-turn crashes show higher non-motorist hit rates than left turns
• RTOR crashes show dramatically elevated non-motorist hit rates
• Non-motorist crashes are 2–3× more likely to result in serious injury or death

---

Spatial Mapping & Visualization Outputs

• Static crash point maps
• County-wide right-turn-only heatmaps
• Animated year-by-year heatmaps (2014–2024)
• Intersection-level danger maps
• RTOR-specific non-motorist intersection maps
• Hoverable intersection maps with:
  • Crash counts
  • Serious injuries
  • Fatalities

Output Format

• All maps are exported as:
  • *.html
• All maps are viewable in:
  • Any web browser
  • GitHub Pages

---

Severity-Weighted Heatmaps

Crash Weighting System

• 1 point = base crash
• +4 points = serious injury
• +10 points = fatality

Purpose

• Ensures fatal clusters visually dominate risk zones
• Prevents low-severity clutter from masking lethal intersections

---

Available Demonstrations

• Animated right-turn crash heatmap
• RTOR non-motorist intersection danger map

---

Python Libraries Used

• pandas
• numpy
• geopandas
• matplotlib
• folium
• branca
• glob
• os

---

Research Applications

• Vision Zero safety audits
• RTOR policy evaluation
• Intersection redesign decisions
• Smart city traffic mitigation planning
• Pedestrian & cyclist safety studies

---

Limitations

• Near-miss events are not recorded in PennDOT data
• RTOR identification is inferred from vehicle movement codes
• Coordinate accuracy varies by year
• Pedestrian signal phase data is not available

---

Author

• Developed by: Tianshu Huang
• Affiliation: Carnegie Mellon University
• Focus Area: Smart Cities / Transportation Safety Analysis