Dry Driver was inspired by the recent flooding in Houston that left thousands stranded and with little knowledge about where flooding was actually happening.
What it does
Our webapp lets drivers input their destination and start point and displays a route along with all flooding points that they might encounter along their route.
How I built it
We used .NET to scrape the data from Hounston's 311 and 911 data, inserted it in a database, and used Node.js and MongoDB Geospacial queries to search along the route we get from Google's Directions API. All of this is then displayed on a map in the browser.
Challenges I ran into
We had never interfaced MongoDB with .NET, so correctly adding all of our data was quite time consuming. We also had problems with our algorithm that searches along the route to find the flood events.
Accomplishments that I'm proud of
We have a working webapp that can take (some) addresses and find the route and any flooding events in the vicinity. It's not perfect, but it is (we think) an impressive marriage of data and logic accomplished in 24 hours.
What I learned
We learned how to split up tasks and make sure everyone on the team had something to work on. We also learned a lot about how to query MongoDB for spacial data and setting our models correctly.
What's next for Dry Driver
We plan to add a few key features to Dry Driver in the coming weeks:
- Fixing the RouteBoxer so all routes work properly
- Create an algorithm to reroute around flood incidents
- Rank data sources properly
Try it out yourself!
Check out http://drydriver.pedelen.com/ and put these origins and destinations in to the boxes to see flood data around Houston:
Destinations: 1500 McKinney St R2, Houston, TX 77010 / 3118 Green St Houston, TX 77020 / 501 Crawford St Houston, TX 77002
Origin: 410 Pierce St, Houston, TX 77002