Form 990

Learn More

The importance of collaboration on 990 data

The information generated from the 990 datasets is absolutely fundamental to our understanding of the US nonprofit sector. They contain a wealth of data to explore and analyze, and can be connected and combined with a vast array of other datasets to further increase their value.

A significant amount of work around 990 data is already underway. Researchers, analysts, and sector leaders rely on information extracted from these datasets to inform fundraising activities, analysis, planning, and so much more.

As the sector strives to continually improve and evolve, consolidating our approach and working collaboratively will draw considerable benefits for the entire sector. We can ensure there is less duplication of efforts, open the door to innovation, generate new insights, and lower the barriers to entry.

Click here if you’re ready to join!

Data and resources available in the project space

Here is an overview of some of the key assets and resources available in the project space on the platform:

Datasets

• 990 Efile datasets from 2009-2020

Visualizations

• A Power BI Dashboard, with Python code that generated the cleaned data, to get insight into Form 990 data trends.

• A Tableau Dashboard, displaying information about U.S. nonprofit organizations by location and NTEE classification

  Code

Our code lives in Github repositories, as follows:

• A demo web app for making a subset of Form 990 (as of 2016) data searchable. Clone the repo to see the code, or visit the live version!
• Nonprofit social media data, including websites and Twitter handles, with code to show how to scrape these data from the web.
• Clustering & Visualization of Program Data, using K-Means, in R
• Clustering & Visualization of Text Data, using a variety of techniques, iPython, combining both Form 990 and social media data.
• A financial analysis of form 990 data using Python.
• A cluster analysis of a collection of the work done by DataKind volunteers on harnessing Form 990 data for improved philanthropic giving.
• House Mapper, a project to map poverty in developing countries by identifying household roofing materials, in Python.

User-contributed repositories:

• Data cleaning files and scripts for 990 data Collection of tools for creating, wrangling, and sharing data and analysis on the nonprofit sector using open data sources. Created by Lecy.
• Training data and an initial ML model for determining matching SDG categories based on project descriptions. Created by Nick Ryan.
• Form 990 Organization Browser app. Created by DataKind.

Additional resources:

• Report: DataKind_Giving Tuesday 2018 Insight Report (PDF format).
• Data dictionary: Datakind GivingTuesday Datadive 2017: Guide to interpreting GlobalGiving’s donation data.
• Project information: The Project Pad for the 2017 DataDive on the Open 990 Data - contains additional links and background information on the projects and analysis done during the DataDive.