How experts think we can make open data more, well, open.
Open data has the potential to improve the economy, environment and our society, but there’s plenty of room for improvement before it is actually able to achieve those changes.
While the Government is keen on releasing as many datasets as possible under the guidance of Cabinet Office open data director Paul Maltby, obstacles include a lack of common standards and questions over privacy.
At a TechUK roundtable, Maltby was joined by a panel of experts to discuss the topic.
CBR was in the audience to hear their recommendations on how to tackle the issues. Here’s the five main suggestions they made.
Agree common standards
Jeni Tennison, technical director at the Open Data Institute:
"[Public bodies] all publish largely with different schemas, different structures for the data, they interpret what they need to publish in different ways. Consistency across local authorities is a problem.
"Getting consistency around data will be important. The National Information Infrastructure [a Government initiative to collect and publish open data] will help with that as well, specifying schemas and standards around open data."
She also points out the need for government to commit to the long-term release of datasets.
"There needs to be some guarantees around the publication of important data that startups, business and government can use that stretch beyond one year pilots," she says.
"Something that is written into contracts or legislation that says ‘we’re going to be publishing this data in this format for the next five years’. That’ is needed for people to feel like they can invest in building systems around that data."
In addition, Ian Makgill, MD of government spending data analysis firm Spend Network believes the Government should become more responsive to businesses’ requests for information.
"One of the things that is really important is to somehow beef up the Freedom of Information request process whereby if you request something that [should be] in the National Information Infrastructure and someone hasn’t published yet, you can get an answer in a month.
"At the moment it takes 13 months to get a judgement on someone not publishing their spend data. That’s not dependable enough."
Make data better by using it
Paul Maltby, open data director, Cabinet Office
"The maxim we have is to get the data out as soon as we can, making it open."
Chris Royles, principal sales consultant at Oracle, questions whether some data is too poor quality to rely on, but Makgill says:
"The only way to make it better is to use it. If we say that’s not good enough then nothing’s going to get better."
Use data to answer business questions
Harvey Lewis, analytics research director at Deloitte, believes the focus should be on the business goal you want to achieve, not the data itself.
"Let’s keep having the discussion about open data but let’s not forget we can now really start to use it to answer some cogent business questions.
"Linking different data values allows you to put different datasets together and that provides all sorts of insights. We’re looking at how open data can completely disrupt the insurance industry because the data published in the UK are enormous.
"It’s about licensing so that the data is free to use and reuse and it’s about making those licenses consistent across government."
Tell people what you’re doing with their data
Matt Davies, EMEA product marketing director, Splunk
"Making sure that citizens understand how data is used, what data is collected and what outcomes the data can achieve is essential. There must be good transparency in publishing the data in the first place."
Strive to end data portals
Most open data is only accessible via portals created by the organisation responsible for releasing it.
This makes it pretty hard to find compared to the far simpler way we can discover information on the web by typing a few search terms into Google.
So our experts think we should try to do something similar for open data.
Jeni Tennison, ODI
"We ought to hold in our long-term vision that one day you will just be able to search in a search box rather than having to go to a specific portal. We should all be thinking about how to get to that sooner rather than later, because it’s simply not sustainable [to rely on portals]."
Ian Makgill, Spend Network
"We did some analysis of tender portals and in the last 5 years there were 2000 tender portals used by the UK Government."
Chris Royles, Oracle
"Government agencies are buying tools to make use of their data. [It’s important to] help consumers of data consume that data very easily without necessarily the degree of expertise required today, without having to develop pipelines for every single new dataset.
"By automating it as possible, organisations could automatically consume new datasets. With a machine process to do that, they can make really good decisions automatically."