Inspiration
Shelters have a hard time managing all of the users that come through their doors. We wanted to help facilitate this process through a centralized platform for all shelters to monitor their capacities and those of nearby shelters to provide the best service to those in need.
What it does
Simple sign up for end-users to enable tracking of shelter resources and needs to be met in the right way.
How I built it
A jQuery and static nginx front-end backed by a strongly typed Go backend for application layer services when PostgREST services on top of our PostgreSQL database do not provide functionality required. Additionally R for machine learning to understand what makes a person at-risk for homelessness. All built in the Amazon Cloud with EC2 and RDS resources.
Challenges I ran into
CORS + Go services clashing with nginx setup resulting in multiple headers overlapping and causing errors in the frontend. Tackled through specific header setup for the go services themselves and a highly permissive postgREST layer.
Accomplishments that I'm proud of
Lots of "first time evers" for the team:
- Go language in general much-less http services for application logic.
- Machine learning for an unfamiliar set of data
- Data Generation for large sets of data with reasonable values (e.g. Lat/Long)
What I learned
Data is not clean, clear, or gathered in a consistent manner when it comes to this subject. Even simple, easily accessible CRUD apps with a strong infrastructure backing could make a big difference.
What's next for ninjacrash
First St. Louis, then the world!
Log in or sign up for Devpost to join the conversation.