Libraries Hacked

Promoting open data and digital prototypes in libraries

6 minute read

For a while many people, including me, have complained about the UK public library reliance on CIPFA for national statistics. The process is:

  1. At the end of each financial year library services gather annual data about their services, financial details, and usage, and send this to CIPFA.
  2. CIPFA validate and aggregate these, and create reports.
  3. CIPFA provide the reports and data to paying subscribers, and for a one-off purchase cost of hundreds of pounds.

There is little available to the public. For English library services in the past the DCMS have funded PDF reports for each library service.

With increasingly open government (to some extent), data portals, and open data communities, this has left public libraries behind other services in providing information about themselves. An embarassing situation given the nature of libraries.

It would be nice to have ‘live’ data on the use of libraries, rather than annual statistics. But as a quick win it would be an achievement to redesign the current approach by modifying Step 1:

  1. Library services collect annual data about themselves, usage of their services, and their users, and publish this as open data.

Libraries complete their data returns by July each year, covering the previous financial year. In September 2017 I approached 17 library services requesting through Freedom of Information the data that had been sent to CIPFA. The main components of the request were:

  1. Did the authority complete the stats?
  2. If they did, I would like a copy of the same data they sent to CIPFA
  3. The data should be in the same spreadsheet format sent to CIPFA. After some research I had found evidence that it is reasonable for FOI requesters to specify data format.
  4. The request mentioned there could be no exemption based on the data being ‘reasonably accessible’ elsewhere. Paying £650 for a complete dataset from CIPFA doesn’t count.

Freedom of Information is an essential process in transparent government, but it isn’t necessarily a friendly one, either for the requester or the organisation. It is a demand for data, and that demand sets the terms of the interaction. Ideally I would have liked to request the data directly from the library service. Even by phoning them, if I weren’t afraid of using phones. But it seemed informal contact would be less likely to be succesful.

A sample request to Bexley can be seen on the ‘What Do They Know’ website.

Some of the responses turned out to be a little annoying.

  • Aberdeenshire sent the data converted to PDF
  • Anglesey never replied
  • Barking and Dagenham chose to use the ‘Future Publication’ exemption saying CIPFA were going to publish in the future. This was overturned after an internal review.
  • Bedford sent the data converted to PDF
  • Bexley claimed the exemption that the data was reasonably available from CIPFA. This was overturned after an internal review.
  • Birmingham claimed the data was still being compiled (odd, given it was well after the deadline). A request to review this was ignored.
  • Blaenau Gwent sent the data converted into 3 separate PDFs. After an internal review they sent the spreadsheet.
  • Bradford replied after 5 months to say they didn’t bother with CIPFA (or FOI it seems).

(You may notice the sample had been selected alphabetically!)

Given that only 17 services were approached, it was a poor success rate. It seemed the biggest barrier to open data in libraries wouldn’t be organisations like CIPFA, it would be libraries. It also became dispiriting chasing these up. FOI has clear processes to follow, but it’s not much fun requesting internal reviews, and following up those that hadn’t responded. Certainly not when coming from an open data community background, where most interactions are between those friendly to the idea of providing public information.

It was interesting to see later that CIPFA had been involved in this. In 2018 I noticed a tweet from Aran Lewis, who had been querying to CIPFA if an individual library service’s data would be available under FOI. CIPFA said no, it wasn’t available via FOI, and included a two page document they had sent to all library services the previous year. This is available in the tweet.

A few key phrases from the document.

CIPFA has been contacted by a number of organisations with regards to the release of information pertaining to the CIPFA stats public library data following a number of Freedom of Information requests made by the same requester.

I’m assuming this was my set of requests in 2017. CIPFA felt it necessary based upon a fairly small sample to warn all services against replying to requests.

If the information being sought is reasonably accessible by other means it is exempt from disclosure […] The fact that a fee is payable does not preclude the use of this exemption.

This is CIPFA claiming that the library service does not have to provide public data, as the public can go to CIPFA for data (£650).

any disclosure that is likely to prejudice CIPFA’s commercial interest can be regarded as sufficient ground for exemption.

CIPFA here are trying to use an FOI ‘trade secret’ exemption to say they would lose competitive advantage if libraries provided public data.

Organisations should therefore exert a degree of caution when responding to any such FOI requests

Some general discouragement from responding to FOI requests.

Public Libraries News took this up on Twitter and described CIPFA as an ‘evil, greedy, and inefficient dinosaur’:

I’m not sure this is fair though. Whatever became of the dinosaurs (and there were some terrific ones), they could not be expected to forsee their demise. CIPFA though are a body that advises and works with the public sector, and will know the direction of travel with open data. Sending warnings to all library services against providing data may cause concern, but it’s fighting against the tide.

CIPFA did provide a full response to Public Libraries News in the middle of this news post, describing their charity status and highly skilled workers. It was interesting to see that only half of Councils with library responsibilities have a CIPFA subscription. The others if they want to see the data must buy it, despite contributing to it.

In November 2018 I contacted CIPFA again, querying if they’d reconsider their antipathy to open data, and even help out with providing the ‘raw’ local authority data as open data. Their own reports and services could remain as commercial interests. They responded with the same lines provided to Public Libraries News, though they added:

We don’t believe the public interest is necessarily well served by providing invalidated data in its raw form.

And this is really where the disagreement is, between CIPFA and any public open data. CIPFA are a registered charity, and to maintain that status what they do will need to be claimed as being in the public interest. But data does not need to be assured by CIPFA for there to be a public interest in it being available. The idea that a small body of non library experts, with limited resource, is a better data validator than the library community itself is absurd.

Those in libraries who have worked on annual library stats will know there are many problems with the data, and know those complexities better than CIPFA do. But those problems don’t ‘matter’ too much, they can be worked on and documented. The main thing is that the data is openly available and can be discussed properly.