Estimating Traffic Disruption Patterns with Volunteered Geographic Information.

Area of Science:

• Geographic Information Systems (GIS)
• Transportation Science
• Urban Planning

Background:

• Accurate traffic forecasting is crucial for policymakers but hindered by expensive data acquisition.
• Increasing road network congestion necessitates cost-effective traffic data solutions.
• Volunteered geographic information (VGI) offers a potential alternative for traffic analysis.

Purpose of the Study:

• To assess the predictive capability of OpenStreetMap (OSM) features for estimating traffic disruptions and volume.
• To determine the effectiveness of different land use categories within OSM for traffic prediction.
• To compare the performance of granular OSM point of interest (POI) data against broader land use categories.

Main Methods:

• Linear regression models were employed using OSM features as predictors.
• Analysis was conducted on 6,500 road network points across 112 regions in Oxfordshire, UK.
• Cross-validation and recursive feature elimination were used to evaluate feature importance and predictive accuracy.

Main Results:

• OSM features alone explained over 50% of the variation in traffic volume and disruptions.
• Specific land use categories demonstrated significant predictive power for traffic patterns.
• Granular OSM POI data yielded superior prediction accuracy compared to broader land use classifications.

Conclusions:

• OpenStreetMap data is a valuable and cost-effective resource for understanding and forecasting traffic conditions.
• Detailed spatial data, such as OSM POIs, enhances the accuracy of transportation and land use studies.
• The findings support the integration of VGI into traffic management and policy-making frameworks.