A search and summary application for traffic events detection based on Twitter data

As a form of social media, Twitter records real life events in our cities as they happen. Huge numbers of tweets under the heading of transportation or metro are published every day. This paper presents an application for Traffic Events Detection and Summary (TEDS) based on mining representative terms from the tweets posted when anomalies occur. The proposed ensemble application contains an efficient TEDS search engine with multiple indexing, ranking, and scoring schemes. Spatio-temporal analysis and a novel wavelet analysis model are applied for traffic event detection. This application could benefit both drivers and transportation authorities. Users can search transportation status and analyze traffic events in specific locations of interest. Utilizing the proposed signal processing technology, we demonstrate the system's effectiveness by examining traffic and metro travel in the Washington D.C. area. As the collaboration between a citizen's life and social media becomes ever greater, this could have a significant impact on the prediction of traffic flow, travel selection, and other city computing functions.