Team Flare comprised of four PA staff members.
We decided as a team to tackle the Spot the Fire 2.0 challenge.

Creating the idea:
The team added each of the bullet points on a white board and designed ideas for completing each task. We then selected a couple of points that we realistically could complete within the time frame provided.

Once we were happy with the selected items:

• Report the Fire
• Verify and screen citizen fire report
• Emergency Social networking

We then proceeded to breakdown the work needed for each task, designing the architecture on the white board and bouncing ideas on how to make the work more efficient and easy to design considering the time limit of one day.

Once this was completed the tasks were decided up between the members.

The biggest challenged faced was piecing together the different parts once they were completed.

We built a responsive web app using HTML 5 CSS 3 and Javascript. The web app was hosted in GCP app engine. We utilised GCP cloud functions to pick up when an image was uploaded to GCP bucket via the web app. Using TensorFlow ML we scanned the image for a fire and give a confidence score on the fire in that area and we used the meta data on the image to grab the geolocation of where the fire is. Then using Firebase cloud messaging we can alert users in that geolocation that there is a fire.

App Engine

Bitquery

Firebase Cloud messaging

Google cloud buckets

Cloud functions

html 5 css 3 javascript front end

https://github.com/MattMoons/TeamFlare