This project investigates how philanthropic foundations describe and frame their successes. By analyzing annual reports and other public documents across the nonprofit sector, we aim to infer an embedding space—a kind of "map"—that reveals the underlying dimensions of performance and enables systematic comparisons between organizations. The student will help collect, clean, and analyze this data to better understand how success is communicated and measured.
How do philanthropic foundations talk about success? Unlike corporations that rely on financial metrics, nonprofits and foundations often use narrative reports, mission-driven outcomes, and social impact language to describe their performance. Yet there's no clear standard for what counts as success—or how it should be measured.
This project explores how different organizations frame their achievements in publicly available reports, grant descriptions, and IRS filings. The central question is: Can we build a shared “embedding space” that captures the underlying dimensions of performance in the nonprofit world? If so, we could enable meaningful comparisons between organizations operating in different domains—like education, health, environment, or social justice—and provide insight into how they set and evaluate their goals.
The USOAR student will work with a small research team to gather and analyze text from a large corpus of nonprofit documents. They will help identify patterns in how success is framed, explore relationships between language and organizational characteristics, and contribute to building a model that places organizations in a shared performance landscape.
This project is part of a broader initiative at the Connected Data Hub at UVA’s School of Data Science, which uses data science to illuminate how institutions describe, justify, and pursue social impact. It blends tools from natural language processing, social science, network science, and public policy.
~ Curiosity about social impact, foundation networks, and non-profit data
~ Basic Python and data analysis
~ Interest in nonprofit work, storytelling, or public communication
~ Text analysis and natural language processing (NLP) techniques
~ How to work with real-world data on nonprofits and philanthropy
~ Basic modeling and embedding methods for comparing documents
~ Theories of organizational behavior and narrative framing