Spatiotemporal Analysis of Traffic Accidents Hotspots Using Twitter Data: The Case of Quezon City

The surge in urban vehicular traffic volume over the past decade has led to an uptick of traffic accidents in busy streets and thoroughfares. These accidents resulted in fatalities, damages to properties, and economic losses. Despite its huge impact on our livelihood, a meaningful spatiotemporal analysis of traffic accident hotspots in urban cities in the Philippines, like Quezon City, remains scarce until now. An additional constraint to performing such analysis is the inaccessibility of relevant data collected by concerned government agencies. To address this issue, this study aims to identify locations where traffic accidents mostly occur (hotspots) in Quezon City, and observe their temporal behavior for a 27-week period using publicly available data gathered from the official Twitter account of the Metro Manila Development Authority (MMDA). Accident locations were extracted from each tweet using natural language processing (NLP) techniques and were subjected to a set of spatial statistics to locate and map accident hotspots. Our analyses show that there is significant spatial clustering of traffic accident locations in Quezon City for the 27-week time series obtained. A stable hotspot was detected along EDSA North Avenue, a disappearing hotspot was found along Commonwealth Litex, and an emerging hotspot was observed along CP Garcia Avenue. The method that we proposed in this study can help promote a data-driven approach to policy making, and encourage the use of publicly available data from social media platforms to uncover insights about vehicular traffic accidents in real-time.