Incydent Here Map

Inspiration

It's not difficult to imagine the profound difficulties that the future may bring. With a warming climate, each year brings the possibility that inadequately maintained utilities will spark the next firestorm and devastate the next or same community. With uneven policy and inadequate tracing, transmissible diseases that escalate to pandemic status with a devastating human and economic toll may become an uncommon and unwelcome future. Small incidents begin to spiral with exponential force.

Incydent rallies the power of the crowd - everyday citizens - to report incidents within their community like dangerous power lines with an act as small as a tweet. This sets off an avalanche of processing and response formalized into a HERE map spatial UI.

What it does

Using AI to automatically parse and classify a stream of citizen created, location based, media rich reports (possibly through twitter using hashtag #incydent), Incydent allows an under-resourced municipal staff to triage the appropriate response with a next generation interface based on HERE maps and data layers that include feature rich details useful to analysts and responders. Incydent uses the Khanboard UI - a hierarchical kanban board - to integrate and organize multiple organizations and levels of organization all with path tracing of the incident report.

By centralizing and spatially centering an integrated response, fewer resoruces can tend to problems before they spiral out of control - literally preventing sparks from becoming flames.

How I built it

Incydent relies on many layers of technology. At the foundation is the data stream that generates, processes, and labels reports for storage into a real time database using standard backend server technologies. This information is spatially indexed and inserted within the Khantext, a spatial map that represents the problem space of the system's Khanboard - a hierarchical Kanban board built using javascript.

Municipal information was downloaded from Here's Data HUB for insertion into a runtime created Here map with data layer information.

Challenges I ran into

There is so much that can be done using Here's information and I believe I've only touched the surfaces. Data layer information can inform the ML classification of the incydent. That same information can be used in forming an active response. The big challenge is incorporating as much as I'd like.

Accomplishments that I'm proud of

Starting to scrape the surface of what is possible.

What I learned

I learned the power of Here's data layer information. I also learned about the flexibility and future customizability of their visual map system that uses Tangram's open source framework.

What's next for Incydent Here

Making this a production system and finding ways to test it.

Built With

• firebase
• here
• javascript
• ml
• node.js
• python
• tensorflow