Habitat for Humanity Restore: Developing and Implementing a Translation Tool for use in Habitat for Humanity ReStore Locations

Project Description

Language barriers in Habitat for Humanity’s ReStore second-hand stores prevent communication between employees, volunteers, and non-English speaking customers. Through research and piloting in the Worcester and Ashland ReStores, we established that Spanish, Portuguese, and Persian are key languages and built a web-based translation solution with Microsoft Translator, OpenAI’s Whisper, and GPT-4o-mini. The solution was deployed on Amazon Web Service (AWS) for scalability, at an estimated $15 monthly in operating costs, and was successfully piloted in real customer interactions. Our project offers a low-cost, replicable strategy for nonprofits to enhance access through translation technology.

Student Researchers

Jenna Goodchild, Kayla Kiely, Christian Reynolds

Project Goal and Objectives

The goal of this project is to assist Habitat for Humanity ReStore staff and volunteers to effectively communicate with customers by identifying and providing viable translation technologies. To accomplish this goal, we had three primary objectives. First, we gauged the available resources from both Habitat for Humanity and their ReStore locations. We analyzed resources such as the budget and technology provided for the project. Next, we determined the translation needs of the ReStore communities. This includes the needs of the customers, volunteers, and employees of both the Worcester and Ashland ReStore locations. Finally, we assessed and tested potential translation tools. This was a two-pronged objective, where we first analyzed which tools have the most potential, and then tested those selected tools with a website we created at the Worcester and Ashland ReStore locations.

Final Report and Deliverables