Traffic Incident Management Crowdsourcing Examples

The following table lists organizations’ use of crowdsourced data (by source) to reinforce aspects of traffic incident management (TIM). Please share your organization’s TIM crowdsourcing example with any of the Federal Highway Administration (FHWA) Everyday Counts Round Six (EDC-6) Crowdsourcing for Advancing Operations innovation coleads listed under Contact Us.

Activity	Agency	Data	Brief Description and Benefit
Incident Detection	Iowa DOT	Waze® INRIX® Twitter™	Traffic management centers (TMC) receive data from multiple sources for detecting incidents on roadways where traditional sensor detection coverage is sparse. In addition to expanding coverage, crowdsourced data has identified incidents 9.8 minutes faster on average than traditional collection methods.
Incident Detection	Florida DOT	Waze	Combines Waze data with computer-aided dispatch (CAD) system data to decrease the response time when crashes and unplanned road closures occur. Data collected from Waze are filtered for roadway crashes and then pushed to each district. Operators confirm the incident and can import the incident event to the advanced traffic management system (ATMS) system rather than manually entering the crash, enabling a quicker response
Incident Detection	Connecticut DOT	Waze HERE® CTroads	Uses crowdsourced data for rapid detection of new potential incidents, improving response times. Newly reported events from Waze are verified using roadside cameras then imported into the State’s CRESCENT traffic incident management system using prepopulated fields. Where camera coverage is unavailable, the Highway Operations Center (HOC) operators may call the Connecticut State Police (CSP) directly or reference CAD data for verification. This streamlined verification process further supports a rapid response.
Incident Detection After Action Review	Kentucky Transportation Cabinet (KYTC)	Waze HERE	Given the high frequencies of Waze reports, KYTC developed an automated process to filter data and send alerts only to TMCs when certain reliability and speed thresholds are met. The automated process also generates an “After Action Dashboard” report. This dashboard combines traffic speed and Waze incident reports in the same graph to provide a holistic representation of the effects of the incident at a glance.
Incident Detection End-of-Queue Detection	Indiana DOT	INRIX®	The agency uses crowdsourced data to identify where quick slowdowns in traffic may indicate the likelihood of an incident. The agency has created a tool to identify when traffic speed changes significantly, as well as the length of vehicle queues. With that information, the agency can alert drivers and take other proactive measures to reduce end-of-queue, rear-end collisions.
Predict High Incident Areas	United States DOT	Waze	The USDOT’s Volpe Center trained a machine learning model to identify high frequency incident areas based on Waze crowdsourced data. The model allows problem areas and highâ€‘risk events to be swiftly identified and addressed.