Data is extremely valuable. Many people were surprised when Microsoft offered to buy LinkedIn earlier this year for US$26 billion. But as Steven Wilson, VP of Constellation Research pointed out, LinkedIn’s user data on 450 million registered members can drive many insights into Microsoft’s core customer base, justifying the hefty price tag.
Fortunately, with over 9.4 thousand discoverable datasets currently published on data.gov.au alone, Australia’s public sector is rich in data. But as governments release more and more data, people in government and industry are grappling with how data can best be used, with associated privacy issues a central concern.
Not enough conversation has gone into what hasn’t worked in open data, says David Eaves from the Kennedy School of Government.
For example in Canada, one initiative published Parliamentary debates in a newsfeed, interweaving Parliamentary transcripts with background contextual information and photographs. By providing users with information in a much more useful and user-friendly way, this is displacing the official Hansard publications, with Eaves stating that official staffers prefer this service. But Eaves asks: “Is this sustainable? I don’t think so.”
Savings must be demonstrable in order to get traction with a minister. Unless you can quantify the dollars saved in departmental budgets, new initiatives won’t receive funding. But if you can do that however, the resources needed to get an initiative off the ground may come your way.
A simple way for public servants to calculate the return on investment of a new initiative is, for example, to see how much time it will save employees. A program that saves an hour of time a week per employee added up over one year can translate into savings that justify the spend. While this is not easy, as demonstrated by the difficulties commonwealth departments had in coming up with tangible productivity savings during the efficiency dividend drive in the last few years, other tips Eaves offers included urging participants to look for internal uses of data. “It will be easier for you to identify and to understand the savings involved” and to receive feedback from the programs being implemented.
Like public libraries of yesteryear
To re-emphasise a point made at the beginning, governments have a wealth of data. As Eaves observed, most open data portals come from governments and the primary users of these data sets are from government. It is necessary for governments to release as much data as possible — even in a raw form.
“We are in a learn and teach phase” states Eaves. Open data is like the public libraries of yesteryear. When we first invented the printing press, we stocked libraries before many people knew how to read, crossing all manner of topics and subject matter, including controversial subjects like political ideologies. It is important to release data now. Otherwise people will get used to working with and become experts in the datasets available in other countries like the USA or the UK.
Some parts of government are taking advantage of leveraging open data to provide services to citizens. Eaves enthusiastically spruiks the open data initiatives in NSW, noting the states was a good example of leveraging open data that few governments are doing well. FuelCheck, provides consumers with real-time information about fuel prices at every service station across NSW. It requires service station operators to submit their fuel prices to FuelCheck. Motorists can search by fuel type for the cheapest fuel available and get directions to the station. A consumer that finds the service center’s actual prices don’t match the ones advertised on FuelCheck can make a complaint by simply snapping a photo of the service station price on their phone, and uploading it to the tool.
Eaves makes a cautionary note — “Data is political … use it responsibly”. The potential for data exploitation has been shown through the redistricting scandals in the United States. In Canada, they have recently made their census voluntary, basically “destroying” the data. Eaves states that this was because data had given the government too much granular evidence about who was underprivileged and where they were located.
Census and securo-crats
In Australia, debate about open data and especially the Census has focused more on privacy concerns. Australia’s Public Data Policy requires non-sensitive data to be open by default. But as many presenters noted, from Mike Bracken of the UK GDS, and chief data officers from the ACT and NSW governments, as well as roundtable participants, government should share more data than currently. While some data is sensitive and needs protecting, the majority of de-identified data is okay to go into the public sphere. Perceived, rather than legitimate, privacy issues and ‘securo-crats’ are presenting barriers.
Another presenter who talked in-depth about open data issues at the conference was Gregory Touhill, a deputy assistant secretary at the US Department of Homeland Security, when presenting about his agency’s cybersecurity program. DHS, which works with over 3000 entities at the federal to local level as well as international partners has an open data by default policy. The department aims to rapidly declassify data (within 24 hours) to share with its partners. Touhill noted that while some pockets of government data require classification, the majority of it can be declassified. Despite its ambitious program, the department has been able to successfully balance open data and privacy issues and as Touhill noted, unusually for government, was awarded a civil liberties award.
Initially people in the department were uncomfortable with this policy, but this is a leadership issue — leaders must instigate the changes necessary to drive an open data mindset. Otherwise, as Touhill summed up, “data that is not available is not useful.”