Open Misinformation and the Illusion of Transparency
====================================================

  | This post is purposely made on the site eris.okfn.org, `Eris`_ being the Greek
  | goddess of chaos and the patron of `Discordianism`_. In no way does it reflect
  | the official position of the `Open Knowledge Foundation`_ on the subject, but
  | comments or even an official response from the OKF or the `UK Government`_
  | would be very welcome. To this end, breaking with usual practice here , a
  | comment facility has been included for this post only.

-----

If you believe the press releases, a significant milestone in governemnt
transparency has been achieved with the release of all sorts of 
`departmental spending data`_ by the `UK Government`_. This sounds really
good. Definitely a coup for accountability or at least for public relations.
If you go to the clearing house at http://data.gov.uk/ you will see
(as at the time of writing), right in the middle of the web page a link
that claims it will show you "how your money is spent". Following that 
link takes you to what looks like a searchable ledger.

So what exactly is this data? Where does it come from? The answers to these
questions are far from obvious without quite a bit of detective work. It
turns out, following the links to the http://openspending.org/ site and 
working by process of elimination that this is a consolidated list of all
the spending over 25,000 GBP (though it also includes a lot of spending
under that value) compiled from individual reports by various departments.
According to both `OpenSpending`_ and `CKAN`_ ^H^H^H^H `TheDataHub`_, the source of
the information is ultimately the `Number 10 Transparency Site`_, the 
links to which very unhelpfully give a 404 error.

All this is far from obvious, and brings me to my first point:

  * In order to be credible, information provided ostensibly by the 
    government should have some obvious way of tracing it to the source,
    and figuring out how it was processed,
    not merely referring to a third-party website that got the data from
    who knows where and did who knows what with it.

This is not to say that the OKF (who is behind the `OpenSpending`_ project) is
doing something dodgy -- they're not, or at least not purposefully. The problem
is that it is difficult to tell what is being done and the emphasis is more 
on providing impressive looking demonstrations than on making available
data and tools that can be used for any real analysis. The evidence for
this is that providing even "soft" human-readable provenance information
that will help understand what they have done is merely an afterthought at
best.

In this case, as a condition for endorsing `OpenSpending`_ the `UK Government`_
should have *required* that they make clear and obvious where the data
has come from and how it was processed.

So what happens when we try to actually use the http://openspending.org/
service for some real analysis -- bearing in mind that this service is linked
from the `UK Government`_ website, without any kind of disclaimer or caveat
so we can naively assume that the data is accurate or complete or at least
useful in some way. Let's try this question: how much has the UK Government
paid OKF for their work on this stuff. Some querying and clicking around
eventually gets us to:

    http://openspending.org/entity/open-knowledge-foundation-limited

A total of 70k seems a mite low for an organisation who has had (according
to what can be publicly gleaned from the mailing lists and software commit
logs) up to a half dozen staff working on various UK government contracts
for the past couple of years. Maybe it is correct, but I am pretty sure it 
isn't. Without other data that this can be checked against, we just can't
tell.

Some more navigating around, and some more process-of-elimination work,
we eventually come to this statement:

    This data is known to contain errors, and corrections are invited from
    the public. Errors may derive from the original source data or have been
    introduced during processing. 

    http://openspending.org/dataset/ukdepartments

Well, that's an understantement. There is a difference between *errors* in
the data and data that is *absent*. An error might reasonably be corrected
by "the public" in some circumstances but missing data cannot be.
Quite apart from that, even this anaemic disclaimer is pretty hard to find
and it would appear that this information is being presented as being 
accurate or complete or useful in some way by the `UK Government`_ when it
in fact is not. And this is the second point:

    * Whilst it is laudable to publish what information you can, do not
      attempt to pass off incomplete and inaccurate data as an answer to
      the question of "how your money is spent". Be honest and up front
      about the shortcomings otherwise it has the appearance of simply
      a marketing exercise calculated to create the illusion of 
      transparency.

I don't doubt that Number 10 and the UK Government are doing their best,
that this represents some of the better data at their disposal that can 
be made public.  A lot of effort by many people have gone into making this
happen, and it is progress. The problem is not that this partial data is being
published but that it is being represented as something it is not.

We need to do better.

-----

At `danbri`_ 's suggestion, I have explicitly requested OKF's side of
transactions with the UK government:

    http://lists.okfn.org/pipermail/okfn-discuss/2011-July/007875.html

The reply from OKF,

    http://lists.okfn.org/pipermail/okfn-discuss/2011-July/007890.html

Looking at the data from the Cabinet Office (COI), at this time OKF has
billed about 500k GBP and has been paid 400k GBP. Even assuming that 
the most recent data has not yet been released from the government side,
what they *have* reported, two transactions for 42000 and 28600 do not
match anything in OKF's books... Curioser and curioser...

-----

.. _Eris: http://en.wikipedia.org/wiki/Eris_%28mythology%29
.. _Discordianism: http://en.wikipedia.org/wiki/Discordianism
.. _Open Knowledge Foundation: http://okfn.org/
.. _UK Government: http://data.gov.uk/
.. _departmental spending data: http://ckan.net/package/ukgov-25k-spending
.. _Number 10 Transparency Site: http://transparency.number10.gov.uk/money.php
.. _OpenSpending: http://openspending.org/
.. _CKAN: http://ckan.net/
.. _TheDataHub: http://thedatahub.org/
.. _danbri: http://danbri.org/foaf.rdf#danbri