Some thoughts after Open Data Camp 3

Although I often describe myself as an Open Data activist, I am not a “true believer” and if you had told me that Open Data Camp would reach its third event, with strong requests for a fourth, I would have probably dismissed it as wishful thinking.

And instead, the community is alive and well. Going from the about 80 attendees of Open Data Camp 1 to the 93 on the first day alone of Open Data Camp 3 is an incredible reminder that the drive towards transparency and efficiency is still going strong. What lessons can we take from it?

Senior engagement: from the centre to the periphery

One comment I have heard is “the lack of participation from Cabinet Office is a very negative sign”. To put things in perspective, we had the Transparency Team in full swing two years ago in Winchester, and Paul Maltby came to Manchester last year to give one of the unkeynotes. No one this year.

I think this might represent a good thing: the participation of many from the FSA (who kindly co-sponsored the event) and DEFRA is very suggestive of a move of Open Data from the centre to the periphery. Government Departments and agencies are discovering that engaging directly with their data users works well, and this engagement does not necessarily need a central push.

Of course, it would be disingenuous to believe that data can be powerful without some direction and standardisation at the highest level in Government, but community engagement works best on a smaller scale — the scale at which experts in a given area (the environment, food, etc…) meet and discuss.

In this respect, the FSA seems to be following an interesting route. I can’t be grateful enough to Julie Pierce, their Director of Openness, Data and Digital, for attending the camp and pitching a session to discuss in practical terms where their data releases should go. This is incredibly positive engagement from a senior level with the whole community.

A manual on how to do data well

The FSA is showing the way to do Open Data right by following two simple patterns:

1. releasing datasets they have a statutory requirement to collect, such as the hygiene rating data. Whenever a public agency is legally required to collect data which has no data protection implications, then such data should be published with an open licence. The presumption of publication for such data is extremely important.
2. their intention to engage with businesses to encourage them to release data they generate and hold, in order to compile new datasets. Many aspects of this have been discussed at the camp, one key issue being how to reward good behaviours. Some ideas revolve around creating standards, some sort of “ISO certification” for food providers, which they could use to show consumers a drive towards quality.

The second point is extremely important because it shows how to do Open Data whenever there are no legal ways to enforce the data release. This might not work in all contexts (the ODI certificates haven’t been picked up as much as everyone hoped) but it’s a way that should be followed because of its potential to sustain positive feedback loops.

If we cannot follow these routes, we need information rights

Thanks to the incredibly well-thought points raised by Owen Boswarva in his sessions, it is clear to me that there are situations when we should expect more. Open Data cannot develop unless we attach some rights to (at least part of) it.

How and when we do this it’s up to discussion, but Government support and understanding of the benefits is vital. Sure, it is not easy in the British context to mandate that all public data be open with no strings attached, but cases such as the sale of PAF together with Royal Mail should not be repeated (for example, with Land Registry). This requires an understanding of the value of data as an asset.

Feedback, evidence, impact

One of the most interesting problems is how to assess the impact of Open Data. I have never thought that this can only be done in a quantitative manner, but we need to discuss what is really the benefit of releasing data.

A problem has emerged, raised by many data producers: in the current mode of interaction and licensing, it is close to impossible to monitor what data users do once they have downloaded the dataset. API are considered a barrier, so a question remains — and it will possibly one of the most important topics in the next few months: how do we collect feedback and use it to provide evidence?

Stay tuned

Open Data Camp 4 will be soon announced. In the meantime, keep an eye on Broken Toilets issue 3, out in the next few days, which will be about Open Data (with an article by me on some of the issues I have mentioned here).