Abstract
Crowdsourced data can help complement official law enforcement data sources in certain situations. This article compares crowdsourced Waze data with computer-aided dispatch (CAD) data in Florida, U.S., to identify situations where better Waze integration could improve law enforcement response. Waze is a smartphone application that roadway users can use for mapping, navigation, and reporting of traffic incidents, roadway hazards, or other travel-related situations. To the authors’ knowledge, no previous studies have compared crowdsourced and CAD data. One year of Waze and CAD alerts reported to Florida traffic management centers were collected. Buffers of 30 min and 1 mi were used to match Waze and CAD alerts, resulting in 6,147 matched events. These events were analyzed with respect to time of day, regional district, and limited-access roadway to determine when and where Waze or CAD reported the event first and the time differences between when Waze and CAD first reported the event. Based on these analyses, it was found that Waze data can provide earlier notification to law enforcement during late night and early morning hours, in districts with smaller urban areas, and on urban toll roads. Even though the frequency of Waze alerts in rural areas is low, they can provide earlier notification compared with CAD data when they occur. To better understand which features were most important in determining whether Waze or CAD was earlier, an extreme gradient boosting model was developed. This model indicated that improving Waze integration in district 2 (northeast Florida), on State Roads 91 and 821, and between 9:00 p.m. and 12:00 a.m. would likely provide the most benefits. Law enforcement and transportation agencies can use these results to better utilize crowdsourced data on their roadway networks.
Get full access to this article
View all access options for this article.