Spatial regression analysis of traffic crashes in Seoul

Area of Science:

• Urban Planning and Transportation Safety
• Geographic Information Systems (GIS) and Spatial Analysis
• Traffic Engineering and Road Safety

Background:

• Traffic crashes exhibit spatial correlations often overlooked in practical planning.
• A gap exists between academic research on spatial crash analysis and its application by planning agencies.
• Understanding spatial factors is crucial for effective traffic safety interventions.

Purpose of the Study:

• To analyze the spatial distribution and frequency of traffic crashes in Seoul, Korea.
• To identify key spatial variables influencing traffic crash occurrence.
• To compare the effectiveness of different spatial regression models for crash data.

Main Methods:

• Development of a traffic crash database at the traffic analysis zone (TAZ) level for Seoul, 2010.
• Integration of GIS-developed spatial variables into the analysis.
• Estimation and comparison of spatial error models, spatial lag models, and ordinary least squares regression, including geographically weighted regression.

Main Results:

• Roads with speed limits under 30 km/h and a higher proportion of residents under 15 were associated with lower crash frequency.
• Increased resident mobility, higher vehicle-kilometers traveled, and specific access point designs (speed limit differences > 30 km/h) correlated with increased crashes.
• Bus-only center lanes demonstrated the largest positive effect on traffic crash frequency.

Conclusions:

• Spatial analysis reveals complex relationships between urban characteristics and traffic safety.
• Design considerations for road merging and access points are critical for mitigating crash risks.
• The significant safety impact of bus-only lanes necessitates further investigation and design improvements, particularly for pedestrian access to bus stop islands.