The IRS website has an Exempt Organizations Select Check that allows the user to search for organizations that:
Are eligible to receive tax-deductible contributions
Were automatically revoked
Have filed Form 990-N (e-Postcard)
After selecting one of these options, a link will appear to download the entire database of organizations. The databases (plain-...
Resource.org has gathered reports dating back to 2002 and says that they process new data monthly. Bulk data can be pulled from here.
For those who prefer a simple search, the Economic Research Institute (ERI) has a database of forms dating back to 2003.
In at least one case, a journalist on the NICAR-L list reported back that there was a form in the ERI ...
From the about page:
The Sunlight Foundation is a nonprofit, nonpartisan organization that uses the power of the Internet to catalyze greater government openness and transparency, and provides new tools and resources for media and citizens, alike. We are committed to improving access to government information by making it available online, indeed redefining ...
There's also a good list just released from ProPublica starting for 2012 http://projects.propublica.org/nonprofits/
And you can always request a 990 directly from an entity -- it is required to give it to you.
Also, GuideStar aggregates them and allows you to obtain one-at-a-time at no cost. Aggregate data will cost, however.
Carl Malamud of PublicResource.org makes the 990 filings of non-profits available for download also via Wayback Machine (2016) in bulk, as well as extracts of the first two pages of these filing. These are the same filings that GuideStar uses for to build much of their service, though I do not think they use Public Resource as their primary source.
Without being a trained copyright expert, the crux of your endevour lies in your process of aggregating your text corpus:
Over the years, I collected texts from many different internet sources (Blogs, Forums, Amazon, News portals, Project Gutenberg, etc.) by my own, which I aggregated and compiled into various data sets that could be very valuable for the ...
Selected financial information on all entities that file using form 990 with the IRS is public. A variety of datasets can be found in the IRS Statistics of Income (SOI). I would start at this page:
Also the National Center for Charitable Statistics (yes, it exists) provides a decent web interface and data downloads for a wide variety of registered and non-r4egistered orgs. No API that I can see though.
Citizen Audit is a good alternative to guide star.
The search process is less click hungry.
Getting several fiscal years at a time is possible on a unique page.
More fiscal years available for IRS 990. ( generally go until 2005/2006, where guidestar give, in most case, the last 3 years.
API and bulk access are a possibility ( if ...
CharityNavigator.org includes a page for every US registered nonprofit. Even if it isn't a rated charity, there is basic data included via a link to the org's Form 990. If the org is large enough (such that it files the full 990), then you'll find those data points on its Form 990.
It's an interesting question. Salaries are the large expenditure. Will the staff be upset if their salaries are public?
I think transparency in a non-profit is essential, especially for charities. If the staff don't want their salaries public, perhaps share the average or median salary as part of an expenditure breakdown.
If you are interested to increase ...
As of this posting, the PublicResource.org link doesn't work, the NCCS datasets are also paid, you must apply for OpenCorporate dataset use.
Below is a more detailed list including Activity and NTEE codes from an IRS page found though a Data.gov search: