ODI Futures and Open DEFRA

Last week we were honoured to have been invited to speak at the “ODI Futures: Open data and agriculture in business” event at DEFRA, on the topic of open data in agriculture and nutrition, and our 5-minute “flash talk” which rounded off the event, was delivered by Martin Howitt. Here, Martin reflects on the event and expands on some of the themes we covered in a bit more detail.

 

DEFRA Secretary of State Elizabeth Truss opened the ODI Futures event on Wednesday 2nd December by challenging us all to do better with open data in agriculture and nutrition. The big issues in this area are regularly rehearsed – there’ll be 9 billion mouths to feed by 2050, biodiversity is under pressure as never before, climate change is a very real threat to food security, and agriculture is facing major commercial challenges.

The Secretary of State set out the ambitious #opendefra programme and asked specifically for ideas on what more could be done. This was followed by a series of short talks and panel discussion from Alex Coley (Head of Data, DEFRA), John Crawford (Interim CEO Agrimetrics), Robert Allen (co-chair at Agri Tech East), and Liz Carolan (International Development Manager at the ODI).

Following a short break there were flash talks from Rob Russell (Senseye) and Martin Parr (CABI) before I spoke and then the event wrapped up with a short panel discussion.

I want to just expand on the 5 things I asked for during my talk as there were some technical angles that you can’t do justice to in such a short time. So here they are:

  1. More data

It’s great that DEFRA are so ambitious in terms of the number of data sets they want to open up; however I wonder if we still aren’t scratching the surface of what is potentially available, and where they have lead others ought to follow. I’m thinking particularly of datasets held by the private and voluntary sectors, streaming data from increasing numbers of environmental sensors that are being deployed, datasets created by mashing up (and cleaning up) existing open datasets, and anonymised data from the increasing numbers of digital transactions that we are putting online as part of the drive to “digital by default”.

  1. Integration

I said during the talk that I was “bored” of negotiating the release of individual data sets. I want to clarify that a bit – of course it’s important that we engage data owners and get them to understand and sign up to all the implications of releasing their data, and I’m as passionate as ever about it. But I also think that if we are to scale up the extent of open data releases we need to open up some new fronts in the release of data:

  • When public bodies procure new IT systems they ought to make provision for open data releases as part of standard non-functional requirements specifications. This means that, in theory at least, systems don’t get bought unless we can get at the data through standard APIs and that the ramifications of data release are factored into every project. I know at least one council that is looking at doing this already, but government is in a position to accelerate this by mandating it in some form. Failing that, good practice in this area needs to be shared and replicated.
  • There doesn’t seem to be a massive amount of understanding as to how specific application architectures can easily facilitate open data release. This is a blog post all to itself, but in essence particular application designs imply specific approaches to getting data out and enable automation of data releases. Robert Allen touched on this topic in his talk and in my view it’s an area that doesn’t get enough attention.
  1. Open Registers

Linked to the topic of applications architecture is the emerging conversation that seems to be happening at the Government Digital Service (GDS) about canonical registers: these are data sources that are regarded as authoritative on a specific topic; for example the Register of Births and Deaths, the Local Land & Property Gazeteer, or the electoral roll. My ask in this area is that specific open data registers are developed (which may in some cases be a subset of existing closed data) and that this forms part of the NII. I know that there’s some thought traffic between GDS and the ODI on this topic right now so I’m happy this is moving forward in capable hands.

  1. Infrastructure

I won’t labour this point, as it was said pretty well recently by Peter Wells: but the fact is that in 2015 our connectivity isn’t anywhere near where it needs to be, particularly in rural areas.

  1. Funding and focus

Rural data services catapult anyone? I just think that rural areas don’t get a fair crack of the whip compared to their urban counterparts. There’s an opportunity to build an international lead in this area.

Overall I enjoyed the day and it was good to meet such an interesting array of people – I’d like to thank Tom Heath and his team at the ODI for organising it.

My talk and slides are below (sadly not in the same link, so you’ll need to imagine which slides were showing at which times) and the slides are also available via this link.

 

My slides are below but also available via this link

Posted in Events.