Two of our teammates have personal experiences with wildfires: one who has lived all her life in California, and one who was exposed to a fire in his uncle's backyard in the same state. We found the recent wildfires especially troubling and thus decided to focus our efforts on doing what we could with technology.
What it does
CacheTheHeat uses different computer vision algorithms to classify fires from cameras/videos, in particular, those mounted on households for surveillance purposes. It calculates the relative size and rate-of-growth of the fire in order to alert nearby residents if said wildfire may potentially pose a threat. It hosts a database with multiple video sources in order for warnings to be far-reaching and effective.
How we built it
This software detects the sizes of possible wildfires and the rate at which those fires are growing using Computer Vision/OpenCV. The web-application gives a pre-emptive warning (phone alerts) to nearby individuals using Twilio. It has a MongoDB Stitch database of both surveillance-type videos (as in campgrounds, drones, etc.) and neighborhood cameras that can be continually added to, depending on which neighbors/individuals sign the agreement form using DocuSign. We hope this will help creatively deal with wildfires possibly in the future.
Challenges we ran into
Among the difficulties we faced, we had the most trouble with understanding the applications of multiple relevant DocuSign solutions for use within our project as per our individual specifications. For example, our team wasn't sure how we could use something like the text tab to enhance our features within our client's agreement. One other thing we were not fond of was that DocuSign logged us out of the sandbox every few minutes, which was sometimes a pain. Moreover, the development environment sometimes seemed a bit cluttered at a glance, which we discouraged the use of their API. There was a bug in Google Chrome where Authorize.Net (DocuSign's affiliate) could not process payments due to browser-specific misbehavior. This was brought to the attention of DocuSign staff. One more thing that was also unfortunate was that DocuSign's GitHub examples included certain required fields for initializing, however, the description of these fields would be differ between code examples and documentation. For example, "ACCOUNT_ID" might be a synonym for "USERNAME" (not exactly, but same idea).
Why we love DocuSign
Apart from the fact that the mentorship team was amazing and super-helpful, our team noted a few things about their API. Helpful documentation existed on GitHub with up-to-date code examples clearly outlining the dependencies required as well as offering helpful comments. Most importantly, DocuSign contains everything from A-Z for all enterprise signature/contractual document processing needs. We hope to continue hacking with DocuSign in the future.
Accomplishments that we're proud of
We are very happy to have experimented with the power of enterprise solutions in making a difference while hacking for resilience. Wildfires, among the most devastating of natural disasters in the US, have had a huge impact on residents of states such as California. Our team has been working hard to leverage existing residential video footage systems for high-risk wildfire neighborhoods.
What we learned
Our team members learned concepts of various technical and fundamental utility. To list a few such concepts, we include MongoDB, Flask, Django, OpenCV, DocuSign, Fire safety.
What's next for CacheTheHeat.com
Cache the Heat is excited to commercialize this solution with the support of Wharton Risk Center if possible.