Generosity AI Working Group
Generosity AI Working Group

990 Data Infrastructure

Providing a vital industry resource – our platform will offer accessible, curated and standardized US nonprofit tax data.

The 990 Data Infrastructure Project, led by GivingTuesday, is a collaborative initiative that aims to provide the public with centralized access to comprehensive 990 datasets and tools.

+ 5.3 Million Files Added

Our Vision

  • Enhance transparency in the nonprofit sector by making the 990 dataset more accessible and user-friendly.
  • Equip social sector practitioners with vital resources to support data-driven decision-making and research within the sector.
  • Establish the core infrastructure that will allow social sector practitioners to build tools and initiatives for the sector’s benefit.
  • Host an environment for social sector practitioners to collaborate on shared initiatives and research.

What does this really mean?

Raw Data

We provide the sector with raw 990 data via our Data Lake, which holds Indices and raw e-filed 990 data in XML format, and is regularly updated with new IRS releases via an ETL pipeline.

Tools and Assets

We are provide multiple tools and assets to further assist users in manipulating the 990 dataset, including a mapper, parser, metadata tables and more.

Processed Data

We will release a relational data store serving as a data warehouse that is preloaded with multiple 990 data sets, including cleaned e-filed data, IRS BMF, Pub 78 and the Auto Revocation List.

Enhanced Access

We will continuously upgrade and facilitate access to 990 data through our API and by providing pre-curated data marts based on common needs, including Mission Statements, Programs/Services, Officers/Directors and Grant data.

990 Datasets

This section offers user access to raw 990 files via the 990 Data Lake, with different options based on common user needs.

Access Raw Data Via AWS Account

This is our suggested method of access, all you need is an AWS account.

Instructions & Access

Download Raw Data Via Browser

If you do not have an AWS account, you can download sample indices directly from your browser.

Instructions & Access

Access Raw Data Via Command Line

For programmatic access, please follow the instructions and prompts listed below.

Instructions & Access

Tools & Assets

GivingTuesday is dedicated to providing the sector with useful tools and resources to advance their research. Follow the link below to explore our current tools and to propose new tools you would like to see here.

The 990 Tool and Asset Repository houses vital tools for navigating the 990 dataset, which can be used to extract valuable financial insights, promoting transparency in the sector.

990 Tool & Asset Repository

Project Roadmap

The 990 Data Infrastructure Project is continently evolving to include a wide range of user cases and products.

The current version of the project is one of multiple stages of this initiative that will include online tools, precut data sets, a relational data store, and API access. To increase utility of this infrastructure for our users, we welcome users to the feedback form to log any issues and desired features that can benefit the nonprofit community.

Log your feedback here

Release of raw data via 990 Data Lake

Implementation of IRS Pipeline

Release of Mapping and Parsing Tools

Release of Precut
Data Marts

Release of Relational
Data Store

Data access via API

Supported By

Our Partners

Who We Are

GivingTuesday is a Movement that Unleashes the Power of Radical Generosity Around the World.

GivingTuesday reimagines a world built upon shared humanity and generosity.

Our global network collaborates year-round to inspire generosity around the world, with a common mission to build a world where generosity is part of everyday life.

The GivingTuesday Data Commons is a global network that enables data collaboration in the nonprofit sector. The Commons convenes specialist working groups, conducts research into giving-related behaviours, reveals trends in donations and giving, and shares results and findings among its global community. With over 300 organizational collaborators and 50 global data chapters, the Data Commons is already the largest philanthropic data collaboration ever built.