Sunday, September 8, 2013

Rebranding Open Data: The "Verified" Feed

The Good Housekeeping seal of approval (no endorsement expressed or implied)

The dream is open data. The nightmare is that open data has little or no credibility. Consider this:
  • The stated goal of government is to release as much data as possible to the public: "Government should be transparent...participatory...collaborative." - President Barack Obama, "Transparency and Open Government," Executive MemorandumJanuary 21, 2009
  • The data is supposed to be maximally accessable and usable for the citizenry. - "Government information shall be ensure that data are released to the public in ways that make the data easy to find, accessible, and usable."- President Barack Obama, "Making Open and Machine Readable, the New Default for Government Information," Executive OrderMay 9, 2013
Yet the polls tell us that people overwhelmingly do not trust their government. "Our Jan. 2013 survey found only 26% saying they can trust government always or most of the time, while 73% say they can trust the government only some of the time or never." - Pew Research Center, July 23, 2013 (quote and screenshot)

In essence, government and its data has a credibility problem - a.k.a. a PR (public relations) crisis.

Any expert will tell you that in a PR crisis, trust is regained by doing a few fairly obvious things:
  • Admitting the problem;
  • Taking concrete steps to resolve it;
  • Bringing in a third-party expert to audit and review your actions.
In the case of government and open data, there are a few practicalities that get in the way of following classical PR advice:
  • We still have to get all releasable data sets released - e.g. it's early in the process;
  • Even if we had all available data, it's not yet available in one place such as;
  • It is impractical to send in outside auditors to every agency to call attention to data inconsistencies and errors.
Yet there is one thing we can do right now to both increase the efficiency of open data collection and increase the credibility of this data with the public. That is the concept of the "verified" feed, marked with some seal akin to the "Good Housekeeping Seal of Approval."

The basic idea is to release the dataset in the best shape one can, and then mark it as "Official Government Release" in such a way that it can always be traced to its official source as authentic (both externally and within the dataset code). Then, rather than spend government agency time and money building elaborate applications to display the data, make the feed available to the public in different venues - both on and on any website where interested stakeholders congregate. 

Ideally, at least on, the feed would be opened up in such a way that external reviewers could comment on it, similar to Wikipedia.


  • Situating "verified" government data in one place,
  • Creating an open space for comments, corrections and edits, and
  • Linking back to it from places where interested parties congregate, so that they see it and are motivated to respond,
...the government data feed is de facto repositioned as "just another source of data to be critiqued."

If we were to do this, it is likely that public engagement and trust in the data would be increased, while unnecessary expenses on useless displays of open data would be reduced significantly.

* All opinions my own.