I am staring at Prithviraj Lanka’s screen over a Zoom call and following a Google Analytics graph that spikes at a nearly 90° angle in April of this year.
This isn't Covid 19 Data though.
This is download counts for a tech-for-good app called Strappd.
Strappd a resource hub for at-risk youth and individuals facing homelessness. Within the last three months, the app barreled past the 100,000 download threshold on the Google Play Store and has gained traction through its Apple and web platforms as well.
The surge in downloads is likely due to the pandemic and an insecure economy. Some experts expect a 40% rise in homelessness compared to last year. [1] But Lanka sees that surge as a confirmation that the project he’s spent more than six years working on is getting into the hands that need it the most.
A Six Year Journey
Lanka seriously contemplated what it means to be homeless after a difficult on Denver’s 16th Street in 2014. Lanka told me on his way to work he gave a man a couple of dollars. Later that day, he saw the same man walk out of a liquor store. It was at this point that Lanka wondered if his donation was really helping or hurting. After some research, he found that the issue was much more nuanced than he thought.
His research also revealed that 40-50% of homeless and runaway youth own a cellphone. A lightbulb went off in Lanka’s head. As a software engineer, he realized he could use his skills to connect the unhoused community with services. “Maybe we could create an app and test it out,” he says. He returned to 16th street and began the researching legwork needed to get the app off the ground.
Many users are over the age of 25.
Lanka’s first idea was to build an app to support homeless youth. This group, he thought, would probably have access to a smartphone. He and his team, a combination of freelancers and volunteers, initially plugged away resources designed to connect at-risk youth with services. However, after launching the app in 2017, it became apparent that 80% of their users were aged 25-55. Similarly, a study from the University of Southern California found that 85% of homeless adults use a cell phone daily. [2] With that knowledge, the team changed focus and reached out to more services around Denver. As the app grew, Lanka reached out to other organizations as well. The University of Denver assisted with legal help and Code for Boulder migrated the project to open source. In 2019, he set up the nonprofit Shelter App, Inc in California to handle donations.
Before April 2020, the app was downloaded 35,000 times on the Google Play Store. Now, with 65,000 more users, Lanka hopes to do more. Currently, the most downloads come from LosAngles, Dallas, New York, Chicago, Seattle and San Francisco. Lanka hopes to expand the project to Canada and other U.S. cities. At the same time, user research continues to play a dominant role in the app’s development process. As Lanka shows me his screen, he switches to the Play Store and scrolls through the app’s reviews. He’s responded to many of them, especially the negative ones. One reviewer writes, “I'm sure the intention of this app is wonderful but I can't get the app to work at all.” Other reviewers report that the data is outdated. Lanka asks the users to send feedback and reach out to him through the app’s email.
After surpassing 100,000 downloads, Lanka wants to address his users’ concerns and implement other features that would make it easier for people to interact with the app. One of his ideas is a virtual assistant that would direct people to nearby services. Another feature could make it easier for shelters and other organizations to update their data, including bed availability and operating hours.
Image caption: Los Angeles and Dallas rank as the cities with the most users.
Keeping Up with the Data
Strappd groups its nearly 7,000 resources into five major areas: food, shelter, health, work, and resources such as clothing and hygiene. The closer a service is, the higher it ranks for a user based on their location. If a user chooses not to list their location, the app lists services in the general area ranked by a user-generated kudos system. Listing available resources isn’t its only goal, however. Lanka hopes Strappd will connect individuals directly to service providers. Once the service provider registers, individuals can contact them through the app. Users can also upload photos and their location information to their service provider.
Finding resources is a priority now that many individuals are expected to struggle this year. Currently, volunteers enter data into the app manually but Lanka is hoping to automate that process in the near future. He intends to use data scraping, the process of pulling information from sources through scripts. Some of the datasets he’s considering are from the IRS, HUD, CareerOnestop, IMLS. He also hopes that the volunteers can validate the existing data. “Our goal,” Lanka says, “is to merge all data sets from Hud, the IRS, and CareerOnestop.”
The project is currently looking for Data Quality Assurance engineers to validate data during import and export and Test Engineers to help with creating test cases. The project runs on Python, MongoDB, and Flask. A volunteer can expect to contribute up to 8 hours per week.