A very popular session at Open Data Camp 5 discussed how to measure the benefit of open data.
Session leader Deirdre Lee, the founder and chief executive of Derilinx, which works with the Republic of Ireland and city of Dublin on open data projects, argued that in the early days, people were focused on publishing data sets.
Now that a lot of data is available, debate is moving onto getting people to use the data – and to realise benefit. So, she said, the questions now are: how do you measure the impact of open data, how do you prioritise which data sets to release, and how do you get government departments to embed this into their everyday work?
Participants said it was easy to measure how many people downloaded something; but very hard to get anecdotes about how they were using it.
One suggestion was to put applications or services on top that add perceived value. This may go against the principles of some open data advocates, who don’t think that the people who release open data should build their own services around it.
But the counter argument is that it is more important to make things useful to people than it is to worry about driving downloads.
Another suggestion was just to release more data; and to make sure that more people know about it. Also, to carry out proper benefit evaluations.
Transport for London was in the news this week, after calculating that releasing its timetable and other information as open data was generating something like £130 million of benefit for the capital; via companies building new services with it, or people using it to avoid delays.
Lee agreed that evaluations were vital, and needed to include qualitative as well as quantitative benefits. One session participant suggested this could be done quickly and easily, by asking people to fill in a quick survey or comment about how they were going to use the information when the downloaded it.
Lee said something like this was about to be launched onto Ireland’s data portal. Another participant suggested feedback should be compulsory: if people didn’t value the service enough to rate it, then the government shouldn’t spend taxpayers’ money on it.
Demand driven open data?
Moving on to her second question, Lee said it boiled down to whether data releases should be demand driven. Should departments continue with the current approach, and publish the data they had available, or release datasets they were asked for?
Essentially, the feeling of the session was – both. If organisations though information was going to be useful, they should release it. But if there was a clear demand for a certain type of data, they should see if they could meet it; if they could get a business case together.
Open by default
How does publishing open data become business as usual for public bodies? Participants said the first problem is that lots of government departments don’t know what information they hold. So, the first step is to make sure they have good registers of data assets.
Northern Ireland has a strategy for doing this: and will build terms and conditions into future contracts so information about them can be logged and released.
Several participants argued that departments also need good, solid arguments. For example, releasing data as open data can reduce the number of Freedom of Information Act requests that they receive.
A participant from a government department that has become open by default said it examined FoI requests, to see if data releases could become regular, open data publications. The same participant flagged another big driver: to make sure that departments that released information as open data were also using that data.