Inspiration: 1.6 million New Yorkers have ridden a bike at least once in the past year. At the same time, 5000 New Yorkers have been in a fatal bicycle accident. We believe that if more cyclists were empowered with a safe and reliable map, a lot more people would be cruising than crashing. What it does Our app provides the safest, most reliable route for cyclists to get from point A to B. It empowers cyclists with real-time bike traffic conditions, bike and road accidents to avoid, closed or obstructed bike lanes and local recommendations on the best routes. How I built it:
We used esri's ArcGIS API and Python tools in tandem with a custom routing algorithm of our own design which allowed us to route users to locations while biasing their trip towards nearby bike lanes and avoiding areas which are dangerous for cyclists, like those which have high counts of reported injuries and fatalities. We created algorithms to import data from NYC's Open Data API and process that data, creating a pandas dataframe ingest pipeline to simultaneously use GeoJSON and coordinate systems. GeoPy was used to get exact coordinates and convert them from addresses. We used React JS to build a simple web navigator and Proto to draft a mobile app version to showcase the output of our algorithms.
Challenges I ran into:
We struggled to find a solution which allowed us to do weighted routing, and eventually decided it was most practical (and exciting) to do a totally custom solution - which took quite a bit of time.
Accomplishments that I'm proud of:
We built an interesting travelling-salesman style routing algorithm that works with different types of segments and uses up to date, real information to safely route cyclists around potentially dangerous and life threatening situations. The end product is a highly useful, highly of-the-moment solution, which aids the modern cyclist in navigating under-funded and disjointed cycling infrastructure safely and with ease.
What I learned:
We worked with a number of complex APIs, however our most interesting challenge was designing an algorithm to do highly custom routing - to sometimes follow pre-set paths and other times find its own way.
What's next for Cruze:
We would like to increase our end-user functionality and deploy our product onto both web and mobile, and add in more sources of data and potentially generalize the project into increasing access to other modes of healthy transportation for the modern urbanite.