Homeless shelters are experiencing a shortage of Personal Protective Equipment (PPE) as well as necessities such as food and hygiene products and are in dire need of donations as more and more people living on the streets look to them for resources during the COVID-19 pandemic.


I live in New York City and every day I see homeless people, on sidewalks, on public transportation. I watch as people avoid looking at them and walk past cardboard signs pleading for help. Through volunteering at a women’s shelter, my team member Amulya has seen first hand how short supplies are, even in normal times. With the coronavirus outbreak, the homeless population is uniquely vulnerable, not just in New York City, but around the world. Now, giving people a way to support homeless shelters is even more urgent.

Solution: Bridge4Good

Bridge4Good is a website dedicated to connecting homeless shelters with the resources they need during COVID-19. Donations are collected through online purchases through our website, so donors are not risking exposure to COVID-19 when donating.

Shelters can register with our website and post detailing the exact items and quantities they need so users know exactly where their money goes.

Our platform also helps support local small businesses during these trying times. Local businesses can add web addresses to the shelters' posts indicating where the necessary products are sold online so that donors can purchase items remotely. Bridge4Good will help small businesses generate income during these trying times.

Design + Build

  • Marvel to mock up our front-end design for the website
  • Python (scikit-learn) for machine learning algorithms
  • Flask to deploy the back-end functionality.

Tech Stack

Using logistic regression, we incorporated donor’s location, donation amount, age, and past volunteer work in order to match donors to their best-fit shelter. We also expanded our model to match shelters with a targeted group of donors based on these features.

The data stored in our platform is open-sourced and available via search services such as Whitepages and through US public records.

Future enhancements include additional features related to donors’ past activity: donation to similar causes, past frequency of donation.

Business Model

Revenue Streams:

  • 1.5% transaction fee charged to donors
  • 2% fees charged to small businesses based on the goods sold on the platform

Key Features:

  • Free for shelters and small businesses to get on to the platform
  • Businesses encouraged to provide coupons to donors to encourage repeat visits

Future Plans:

Once the platform is popular, extending the model to charge businesses to get onto the platform as well as implementing preferential ranking and advertising

Next Steps

  • Consider expansion of the platform to iOS and Android users for ease of use
  • Link small businesses and shelters to the platform Enhance ML algorithm:
  • Add features to improve recommendation accuracy
  • Address ethical considerations such as enabling shelters with less donations priority over those that have already received donations
  • Advertise and promote the platform through Instagram and Facebook communities such as @goodnews_movement and @tanksgoodnews.
  • Link social media accounts to donors’ accounts for additional advertising
  • Consider expansion to incorporate mainstream delivery services

Note: the video attached is from Round 2 of the Teens in AI YouthVSCovid Hackathon. The Round 1 video can be found at:

Try It out



ai, flask, machine-learning, marvel, python, scikit-learn

Devpost Software Identifier