Based on feedback on the original post, here is the same data normalized by GDP.
This is a data representation of all non profit programs found on publicly available 990s for the tax year 2023 categorized by their description as a percent of state GDP. These are the aggregate program expenses, not all expenses for all non-profits.
To create the categorizations we generated embeddings for all program descriptions clustered them, then labelled clusters using an LLM.
Tools used: matplotlib, sentence-transformers, umap, hdbscan.
Data provided by [Npo Align](https://npoalign.com/?campaign=reddit) and and https://en.wikipedia.org/wiki/List_of_U.S._states_and_territories_by_GDP
Leave A Reply
Du musst angemeldet sein, um einen Kommentar abzugeben.
1 Kommentar
Based on feedback on the original post, here is the same data normalized by GDP.
This is a data representation of all non profit programs found on publicly available 990s for the tax year 2023 categorized by their description as a percent of state GDP. These are the aggregate program expenses, not all expenses for all non-profits.
To create the categorizations we generated embeddings for all program descriptions clustered them, then labelled clusters using an LLM.
Tools used: matplotlib, sentence-transformers, umap, hdbscan.
Data provided by [Npo Align](https://npoalign.com/?campaign=reddit) and and https://en.wikipedia.org/wiki/List_of_U.S._states_and_territories_by_GDP