Swerve is a computer vision car collision mapping & visualization tool to help insurance agents easily view and interpret collisions. We built a motion tracking system for cars (using Verizon’s API, Mapquest and sensors) to understand car driving behavior at a data-driven level.
Currently, insurance agents have a difficult and slow time doing the “root cause analysis” of car accidents, conducting interviews, seeing lots of camera footage and listening to “he said, she said” stories.
With Swerve, agents can easily see where/when/how a car collision happened in a simple web-based tool - features such as - GPS location, collision time, all cars maximum speed (upon impact), collision history at location (good for time series analysis).
What it does
(1) Swerve uses computer vision for motion tracking of cars (via Verizon’s ThingSpace API).
(2) Swerve aggregates the data images and processes them to create a “pin” on a satellite map (via MapQuest) so an insurance agent can see exactly where an accident happened, automatically.
(3) Swerve displays car collision analytics on a easy-to-view web platform that are relevant to insurance agent needs.
How we built it
Challenges we ran into
(1) Narrowing to right problem - we bounced around a few ideas for the initial hours, but we asked sponsors what were the top problems they faced and then settled on transportation, data mapping and visualization, and insurance agents as our end user.
(2) Verizon APIs - struggled to get started but a Verizon agent helped us understand the documentation for the cameras, sensors, etc. for Swerve.
Accomplishments that we're proud of
(1) Team Rapport - we bonded closely right away and had a fun time working all night together, from wearing blue NEXMO onesies to making coding jokes to learning about cool tech in the transportation industry
(2) Sponsor Connection - we met with all the sponsors about available APIs and tools, and got great insights into what are their best use cases. We settled on Verizon’s after speaking with their team and hearing the benefits of the ThingSpace API platform.
(3) Coding Time Urgency - we were able to code as a team and complete a hack on time.
What we learned
(1) Focus - how to stay focused on one problem, solution and end user
(2) Team Work - how critical it is to work with a diverse group of people and skill sets, from coding to product design to pitching.
(3) Computer Vision - how tracking of objects works and data analysis
(4) Data Visualization & Mapping - how to use the right tools for image display
What's next for Swerve
We are going to test this tool out “in the wild” and talk to insurance agents if it is beneficial for idea validation, and try a beta test. We’re based in SF and NY and will remotely collaborate!