[open-government] The Four "A"s of Open Government Data

Josh Tauberer tauberer at govtrack.us
Sun Feb 12 01:43:54 UTC 2012


Last week the House Committee on House Administration (here in the U.S.)
held a conference on legislative data and transparency. Reynold
Schweickhardt, the committee’s director of technology policy, made an
interesting observation at the start of the day that policy for public
information often is framed in terms of 3 A's:

     accessibility,
     authenticity, and
     accuracy.

I thought about that over the next few hours. They are good principles.
And yet us data geeks so often find ourselves having to start from
scratch explaining why clean data is so important. It seems
contradictory: if accuracy is a concept practitioners in government get,
and if 'clean' is a type of accuracy, then there must be some
communications failure here if we're having a hard time explaining open
data to government agencies. (To be clear, Reynold totally gets it.)

     --------------------------------------------
     TLDR version: Read chapter 5 of my book at:
     http://opengovdata.io/2012-02/page/5/principles-open-government-data
     --------------------------------------------

So I was thinking that morning, what other word do we need to add to
those 3 As to work open data in there? At first I thought about adding
"precision". Precision is one thing we're usually asking for when we ask
for open data. Precision is basically granularity. Compared to say a
PDF, XHTML is more granular because it is explicit about section
boundaries, paragraphs, identifying where in the document the important
things are like names and dollar amounts, etc. (It is more granular with
respect to the meaning of the document, though not its pagination.)

But precision is too narrow. When Congress releases its institutional
spending records, it does so in a PDF. That PDF has high precision ---
it gets down practically to line items. The problem with the PDF is that
it has low accuracy because getting it into a spreadsheet format and
de-duping names introduces errors.

But accuracy is already one of the three As. So what's missing here?

The Association of Computing Machinery’s Recommendation on Open
Government (February 2009) figured this out:

> "Data published by the government should be in formats and approaches
> that promote analysis and reuse of that data."
http://www.acm.org/public-policy/open-government

Not only is it right, but "analysis" starts with the letter A. Plus, in 
order to do any useful analysis on large amounts of information, we need 
automation --- another A word. That is fate if I ever saw it.

Proposing a whole 17 distinct principles of open government data (read 
the chapter!) might be, let's say, overwhelming in any practical 
situation. If we had to do with just four words, maybe these will do:

     accessible,
     authentic,
     accurate, and
     analyzable (using automation, because data is big these days).

Analyzable gives deeper meaning to the other three words. Accuracy is 
too vague alone. You can't measure accuracy in the absence of some 
process. In the computer science world, accuracy is how often something 
comes out right. I think government documents people have considered 
that 'something' to be if a Xerox machine copies enough pixels 
correctly. That's not sufficient for analysis anymore. We can't go 
hiring thousands of interns to read all of the documents governments 
produce. We didn't build computers for nothing.

With analyzable added, the meaning of accuracy is that an *automated 
computer process* will get it right. If someone says a document is 
accurate because it is a scan, I'll say that's what accurate meant in 
the 1960s. If the fourth "A" of government information is analyzable, we 
can redefine accuracy for 2012.

But if you want the full 17 principles, read the rest of the chapter, 
which tackles data quality (accuracy & precision), machine 
processability, and other concepts in more detail. There's also a case 
study on the House disbursements documents, looking at whether and how 
it met the 17 principles:

     http://opengovdata.io/2012-02/page/5/principles-open-government-data

Thanks,

- Josh Tauberer (@JoshData)
- GovTrack.us | POPVOX.com

http://razor.occams.info | www.govtrack.us | www.popvox.com




More information about the open-government mailing list