The idea for GimmeShelter came from the tremendous need to provide a safety net and source of resources for homeless youth. In particular, homeless members of the LGBTQ youth community, which often requires different approaches than traditional methods.
We’ve created a system to connect homeless individuals to the help they need, and provide care organizations with valuable real-time data.
What it does
GimmeShelter will greatly streamline and enhance the data collection process. Rather than handing out long, time-consuming paper questionnaires for people to fill out every time they check into a shelter, our site will ask users a few key questions each time they request a service, populating and updating their profile over time. The questionnaires incrementally gather information on a "need-to-know" basis. For example, if the user has indicated that they are male, they will not be asked if they are pregnant. This will greatly diminish the amount of intake fatigue experienced by both the user and the provider when entering data into HMIS.
This method of data collection will open up a world of possibilities: frequent updates will better enable tracking of individuals throughout their progress, reducing or eliminating the need for paper forms will free employees for other more meaningful tasks, and the production of a richer data set will enhance visualizations and allow machine learning to greater aid prediction and prevention of homelessness in the future.
We expect the user base to consist mostly of homeless youth, who tend to use technology more and own smartphones or other wifi/data connective devices more than adults. While this is certainly a subset of the homeless population, it is definitely a populace that needs attention. Furthermore, if GimmeShelter catches on among the youth, it stands a good chance of being adopted by everyone who can use and benefit from it.
How we built it
GimmeShelter is provisioned in a vagrant development environment, and deployed to Amazon via a bootstrap script. The app features five primary categories:
Gimme Shelter: Walks the user through a series of questions to determine how acute their immediate shelter needs are, and make contact with providers accordingly. If a <18 year old user states that they are experiencing abuse at home, needs a place to stay that night, and has not left their home yet, the system will alert child protective services and have them contact the individual on the phone number that they entered.
Gimme Safety: external MySQL integration w/ Zapier which activates cascading Twilio actions when a client is: a) in imminent danger, b) feeling suicidal, c) reporting cases of domestic abuse and/or violence.
Gimme Jobs: Python scripts scrape Craigslist job postings based on Age and technical skills (checkbox)
Gimme Housing: Python scripts scrape Craigslist housing posts based on Zip Code and show housing options in the area, conforming to a reasonable rent level (1/3 of AMI) determined by collecting a client's AMI.
Gimme Healthcare: Coming soon!
Gimme Benefits: Coming soon!
COC/Provider Features: Each provider will see a dashboard of information concerning their facility and others. Information provided will include facility capacities and trends of shelter entries and exits in the recent and more distant past. A notifications bar at the top of the page delivers important messages, such as homeless teens arriving directly from abusive home situations and other situations that could require additional attention. There is a chat feature for quick communication with other team members and staff members of other facilities, without undue interruption of the workflow.
What's next for GimmeShelter:
We are excited to see what the future holds once we begin obtaining a richer data set on the homeless population! Machine learning and visualizations will become even greater tools for prediction of issues, prescription of solutions, and allocation of resources.