Defra Data Unconference
In late November, Paul Connell and Tom Forth of ODI Leeds had an away day at Defra's Data Unconference at iCentrum in Birmingham. This marks the start of conversations with Defra about helping to unlock the potential and value in their data releases.
Our first impressions were:
- Defra is a huge organisation, with lots of sub-agencies and partners.
- Much of the organisation is doing interesting things with data.
- Lots of individuals and leaders are keen to do more in their area.
- iCentrum has a fantastic coffee machine and great WiFi.
We pitched two events: how to add value using open data and how to build networks with open data. But rather than replaying each session I've pulled out three main points from the day.
Businesses are being built on open data - Defra should celebrate these successes and the innovations behind them.
One of the most encouraging things we saw was a number of people providing great services and running successful businesses that wouldn't be possible without open data. John Murray's work with LiDAR, ASI open data's work on pollution in London parks, and Shoothill's GaugeMap were three great examples.
We think that Defra and its agencies are being too modest by not claiming at least some success for this. If open data powers a product, then the people creating, maintaining, and sharing that open data should take credit for that as well as the people who create the product. Doing so helps the whole data community by taking some pride in providing and maintaining datasets, and showing a commitment to the data community that datasets will not be left to languish on a server somewhere.
Defra is trusted with data - but could achieve more if it released some control.
When we ran #FloodHack16 at ODI Leeds in March 2016 we were impressed with the data on flooding that has become available in the past three years. We have since seen some great things built with it, such as the RAC Foundation's road at risk of flooding tool at #HighwaysHack.
Now we have a chance to achieve even more. The word that sums up how we could do so is "control".
Defra's data on flood risk is not only trusted, but often a legal requirement to support plans for our cities, towns, and countryside. Defra is the authority in this regard. It deserves to be, and must remain so.
But being an authority does not mean that Defra can't release some control. For example, at #FloodHack we were frustrated that we couldn't answer simple questions like "if we change this flood scheme, how will that affect the flood risk here?" or find anyone who could. The reason for this is that while flood risks are released as open data, the models that create them are not. Even internally, where most of the value of open data accrues, people were unsure who built and maintained the models they were using.
Innovations with LiDAR are already improving how we understand flooding. It may be embarrassing to some at the Environment Agency if other people build better models than them, but if that leads to improvements that can be reintegrated into official models, we should all welcome that.
This is a representation of how we saw this working for flood models - this approach could be used for the other domains that 'The Defra Family' looks after.
Defra and its agencies have plans to communicate well internally — but they should keep trying to go to where other people already are.
"Is everyone on Yammer? What about Slack? Who controls the twitter and facebook accounts of our department? Who has permission to publish a blog on our homepage?"
We heard a lot of encouraging things about how Defra and its many agencies communicate now. We heard how they plan to communicate in the future too. But it would have been good to hear more about how they plan to meet and work with more of the people in the UK who are solving similar problems but who are outside of Defra.
In Birmingham, there's ImpactHub, in Leeds there's ODI Leeds and FutureLabs, in Manchester there's Open Data Manchester and Things Manchester, in London there's the Geovation hub. And for each place that I know of, there are hundreds more all around the UK that I don't.
How often do we see Defra people at these places? How many projects are they sponsoring? How would Defra buy a product that they saw there? What does Defra already "know" and how does it share it between itself.
We didn't get any good answers to those questions. But we were happy that there were a few people who started thinking about them, and being more open with your data, how you are working, and on what, is a great way to reduce the friction and communicate in an organisation such as Defra.
ODI Leeds Head of Data