Verification of Crowdsourced Crash Reports Survey 2018

The process by which data for the 2018 Verification of Crowdsourced Crash Reports Survey was collected is decribed below:

1. An algorithm scraped tweets from @Ma3Route in real time and used a machine learning algorithm to detect RTC related tweets. When the algorithm detected a RTC tweet, the tweet was sent to a field coordinator for the project. The field coordinator for the project confirmed whether the tweet reported a crash with sufficient location information and, if so, made a request on the Sendy Ltd platform for a driver to go to the RTC location. Sendy Ltd is a Kenyan on-demand logistics startup that offers a marketplace for deliveries and logistics services using a mobile application that connects clients to motorcycle riders and vehicles.
2. Sendy drivers that were a part of this survey (15 drivers) were alerted of the request. A Sendy driver could choose to accept the request, then would travel to the location.
3. Once a Sendy driver arrived at the RTC location, they filled out a short survey verifying the RTC—if there was no RTC nearby, they were instructed to ask a passer-by if one occurred recently. Sendy drivers were instructed to not go near the crash, and to not interfere with any first responders.