As the world celebrates 25 years since Sir Tim Berners-Lee’s invention of the World Wide Web, two events have shown the potential of his other great passion – open data. On Sunday, I was a judge at Rewired State’s National Hack the Government weekend. On Thursday, I shared a panel to discuss open data in local government at the Ovum Public Sector Enterprise Insights conference. Both events highlighted the positive potential of open data. Both made me question whether our current approach is right. As Sir Tim calls for the world to protect his first famous project, I think it is time to ask how we can ensure the success of his second.
First, a little background. Much has been written about open data’s virtues. The government sees it as a means to increase transparency, spur innovation and improve public services. There is also the question of cold, hard cash: making public sector information open for use has been estimated to be worth over £6 billion to the economy. Happily, the UK is a leader in the field. As of October 2013, more than 10,000 datasets had been released and listed on data.gov.uk – the most comprehensive resource of its type in the world. Thousands more are available through other public sector data portals, such as data.london.gov.uk. Beyond the data stores, the Open Data Institute, the Open Data User Group and the Data Strategy Board have been pioneering in their respective functions.
So why the concern? It is important to acknowledge that releasing open data is not cost-free. For the organisation holding the data, it may require investment in new or updated software, (potentially) the creation and maintenance of a web portal, and the development of APIs. Most importantly, to be of much use, it requires the resource time to annotate the data and provide sufficient support for those who aim to use it. The fact that the Treasury will eventually receive extra tax revenues from data-fueled businesses in VAT or, if they make money, corporation tax, definitely is a long term benefit. But in the meantime, government departments, quangos and local authorities have to run (and aim to balance) their own budgets. To be worth the time and investment, open data must achieve – and be seen to achieve – objectives that genuinely matter both to citizens and government. If it fails to, there is a real risk that future governments could claim that they cannot afford to be open. My concern is that, in the current drive to release as much data as possible, we risk missing those objectives.
Start with Transparency. The idea is that, by releasing datasets about government activity, an army of armchair auditors will be able to spot where and how it deviates from its declared path or budget. Yet there is a danger that, as government releases ever more data via sites such as data.gov.uk, the needles of useful information will increasingly be concealed within giant haystacks of the irrelevant. The term ‘big data’ is often used to describe datasets that barely justify the name. But, in the case of government, big really means big. (HM Revenue & Customs alone reportedly holds over 80 times more data than the British Library.) The blanket release of data in the name of transparency could therefore end up making government not more open but more opaque. Plus, as few individuals or businesses outside government (and sometimes even inside government) have any idea of what data a public sector body holds, they are ignorant about what they could feasibly ask to see. As a result, they have no way of telling whether the data listed on a portal paints the complete picture.
Open data is also intended to spur innovation, as businesses harness public sector information to develop apps and other data products. Again, there is a problem. Imagine that an entrepreneur stands in front of a panel of business angels, seeking investment. The entrepreneur explains their wish to create a business that depends on using just one resource, available from a single supplier, and which has no guarantee of future delivery. You might imagine that the investors would be reluctant to part with their cash to support such a business. This is precisely the case with those that depend on open data. This point is not merely academic. There are local authorities that have let app developers create services using their data, only to withdraw it – or threaten to do so – later. If the open data agenda is being pursued for economic benefit, it is vital that viable and scalable business models can be formed using it.
Finally, open data is seen as a way to improve public services. This is particularly pressing for local authorities, which are expected to see budget cuts of around 40% by 2018. They will surely only be able to cut back so far before they have no choice but to find fundamentally different ways of delivering their services. Collaborating with citizens and civic hackers to use data intelligently and develop new products is one possible way to achieve that. (The dozens of apps created with Transport for London’s data highlights what can be done.) There are developers and groups such as MySociety who are willing and able to use open data to help councils – but they first need to know what problems councils are trying to solve. Regrettably, such two way communication seems rarely to happen. Simply publishing open data to a portal is not enough, as it treats citizens as passive recipients of data.
What might a more strategic approach to open data look like? Several measures could help. To be truly transparent, public sector bodies should declare the key targets they aim to achieve and the measures by which they should be judged. They should then ensure that datasets that track progress against those measures are made easily searchable on the relevant data store. They should also be required to periodically audit and declare all the non-personal datasets they hold, so citizens have visibility of what data they could request, and know what is missing. If we are serious about innovation generating economic returns, government must engage with businesses to identify the datasets of real value and concentrate efforts on ensuring their long-term future provision. Publishing a schedule of when future datasets will be released would also help start-ups and SMEs plan ahead and develop new products. And to improve public services, public sector bodies should be open with citizens about the challenges they face, and positively engage civic hacker groups in creating new solutions. Moreover, with the widespread adoption of smartphones, platforms based on open standards (e.g Open311) should be established to give citizens an automated way to send data to government. This would empower communities to report on and help resolve issues in their local communities, from pot holes to park littering.
Open data is a remarkable resource that can and should be harnessed to make government better. Our ultimate vision must, of course, be that all non-personal datasets are released unless there is a security or privacy reason not to do so. The best way to realise this vision is to ensure that both government and citizens understand its worth. That requires being strategic in our approach to open data, and making sure it actually achieves its goals.
Follow Eddie Copeland on Twitter @EddieACopeland