With Purdah rapidly approaching, and the current ODUG mandate finishing at the end of May, the time is ripe for a reflection on what lies ahead for the Open Data Agenda. I'm fresh from giving a keynote at a conference last week on "Improving public service delivery through Open Data" where I discussed some of the major lessons-learned from the projects funded by the Release of Data Fund: Delivering tangible, real, practical benefits is something on which we now need to focus our attention. There is a huge interest in these outcomes. Open Data is a weapon in the arsenal of those wanting to improve service delivery, and my position is that this is key to where we need to head after the general election.
Open Data is a venture which started and is driven by an effective transparency objective; it has been a fantastic way to help improve the public's scrutiny on the work of Central and Local Government, to analyse performances and expenditure, to question how services are run. This is all good, but this view of Open Data has often resulted in what I call "data littering": dumping data in public repositories without particular care to building structure, defining procedures, or clarifying responsibilities around the release process.
Hence the question: is this model satisfactory, or can we make it better, more dependable, and increasingly capable of delivering tangible benefits? In other words, what lies ahead in the future of Open Data, in terms of release, licensing, and use?
To answer this question, I need to make two points.
First of all: we need to speak about the people who can advance the Open Data agenda.
I believe there are two competing souls within the Open Data community: on one side those who view "Data" as a qualifier of "Open"; on the other those who use "Open" as an adjective for "Data". In other words, activists of the "Open" tradition (Open Source, Open Access, and so on) and "service geeks". In order to speak about the future we need to appreciate this contrast, and work to overcome it by bringing these two souls together to cooperate effectively.
Secondly: we need to understand what happens after the data gets released.
At the end of March we will get the final report on the delivery of the projects funded through the Release of Data Fund, many of which focus on building capacity around data releases. In many cases two side-effects have emerged:
- a more conscious embedding of data in operational choices and everyday tasks
- the creation of data-rich services, or the improvement of pre-existing services through data practice.
We need to monitor uses of Open Data to guide the process of data release and the embedding of data-based intelligence. Back in 2014 we published a collection of Open Data Case Studies with the goal of stimulating this analysis. We need to make such analysis an ongoing effort.
These points made, I believe that four major challenges face us to realise the next stage of the Open Data agenda:
1 - Pushing Open Data where it is not fully embraced
There are sectors where Open Data is not welcome, or effectively fought against, for example in the geospatial and addressing areas. We need to make sure that data created with public money is considered for release as Open Data with high priority. We need also an improved clarity of licensing terms. The OGL is now in its third iteration, and I really welcome Ordnance Survey's move to adopt it for most of its data product; however, "most" is not good enough in many cases. We need to work in order to make sure the OGL is more widely adopted within the public sector.
2 - Achieving genuine (Open) Data by default
Open Data by default is a pressure to release datasets into public repositories; all too often it is then forgotten. This is not good enough. A genuine Open-by-default approach will need to start embedding Open Data in everyday's tasks, moving to a model in which Open Data is not just the final result of a process but the very way processes are run. This is about creating Operational Open Data.
I put the word "Open" in brackets because, after all, this is not just about Open Data, but about a general move to adopt data practice into the way public services are run. This is very important also in the context of shifting our policy-making process to a more evidence-based, data-powered one. As Paul Maltby says in his blog post for NESTA [LINK http://www.nesta.org.uk/blog/new-operating-model-government], "It is not just a case of feeding modern digital tools into our existing policy processes it is about recognising that these technologies have the potential to allow or even require a different operating model for government."
This shift requires a huge cultural change, but without overcoming this challenge we will not be able to enjoy the full benefits of Open Data.
3 - Improving public confidence in Open Data
This is a sore point. Confidence in Open Data is not high across all layers of society. Data leaks, data protection nightmares, initiatives like the undoubtedly well-intended but controversy-ridden care.data, and the conflation of Open and non-Open Data in the press, have shed a bad light over the Open Data agenda. We, as activists and data professionals, need to work together with the public to rebuild that lost confidence and increase trust. We need to do this by showing the benefits that data can bring to society, in terms of business, social action, transparency, and citizens' rights.
4 - Improving (infra)structure around Open Data
Once again, Open Data should be more than just releasing datasets. As I've said before, we need to start working on the operational aspects of data, on embedding data in the decision-making process. This requires a framework around data release that is able to ensure sustainability for the process, accountability for the holders, quality for the data, and clear controls. Data is the infrastructure of our time and age; as we expect other elements of the country's infrastructure - roads, railways, water pipes, electricity lines - to follow effective practices, the same should apply to Data. The vision for a National Information Infrastructure should be doing this.
Many of these challenges will be part of the work a Chief Data Officer for the UK - whom I hope will be appointed soon - is going to have on his or her desk on a daily basis. Nick Halliday, a colleague on the Open Data User Group, has effectively summarised what a CDO should do and I agree with his call.
We will need to work on these challenges if we want to make Open Data even more effective than it has been so far and make sure it positively impacts the business opportunities, the actions of charities, the efficiency of services, and the lives of citizens.