Improving vehicle circulation and fuel efficiency between the cities of Windsor and Detroit would help reduce emissions and allow local traffic to get to their destination more efficiently. Offering trucking companies a data-driven solution to lower idling cost. Allowing key decision-makers of the city to optimize based on insights extracted, and drive a maximum impact on them through the holistic visualization offering.
What it does
Virtua denotes two nodes on the map and calls the Google Maps API to find the most efficient path without taking into consideration the traffic. Virtua could then in generating a time estimation for the travel time between the nodes given using a rolling average from Miovision and City of Windsor censors to achieve real-life prediction. Giving drivers additional insights and better utilize the sensors provided
How we built it
Virtua is built based on a multi-layer visualization architecture, on the lowest layer we have MapBox API providing us with the foundation of the application. We chose MapBox for its efficient integration with deck.gl as well as it's customizable aesthetics. On top of MapBox, we have deck.gl, Uber's visualization 3-D browser library. Deck.gl gives us a diverse set of tools to create the exact traffic interaction/visualization we wish to achieve. Using Deck.gl and Google Pathfinding API, we charted out points, arcs, paths using GeoJson to create the application.
What's next for Virtua
We wish to expand our capacity for accurate prediction beyond real-time. Which is why we began to using feature engineering of existing data for recurrent neural networks. In the upcoming days, we hope to provide a long term prediction capability utilizing LTSM. However, for now, we are proud of what we achieved,