The proliferation of cheap Internet-enabled sensors and services is starting to shed light on aspects of human behavior that hitherto could not have been studied in detail. In particular, the way in which mobility and pro-social behaviors contribute to road safety has remained unexplored. In this work we analyze data from a ride sharing service, Twitter, and accident reports to analyze how mobility, safety mutually influence each other. Road traffic is estimated based on the number of trips passing through a specific area and correlated with the number of reported accidents. Geo-located Twitter activity is also used to identify the popularity of different locations within the city and how it correlated with lower safety records. Major behavioral and temporal patterns are identified and analyzed for each city and differences between cities are highlighted. These results highlight the role social media play as sensors for collective phenomena and could help to better understand the social factors involved in road safety.
Preliminary results from this research were presented at 2016 International on Computational Social Science held on the 24-26 June at Northwestern University, in Evanston, IL, USA.