Learning to love Linked Data

Linked data has been a topic of discussion at successive Open Data Camps. So at Open Data Camp 4 in Cardiff, Jen Williams of Networked Planet whipped through the basics.

Linked Data at Open Data Camp

“When people talk about linked data they are talking about putting it into a statement,” she said. “So in a normal spreadsheet, you have a lot of columns… with linked data you start with an identifier and then go to the column header, the ‘known as’, and then you go to the value.

“And the reason we do this is that it makes every entity in this record addressable, and when we put it onto the internet we can refer to that thing.”

“It’s a bit like a photo. If we want to show somebody a photo, then with a spreadsheet, it’s like saying: ‘here look at this album’. With linked data, because we have identified things, we can direct people to them.”

At previous ODCamps, she added, she had explained this. And people had “got it.” But then, naturally, they’d wanted to know what to do next; and they’d found they didn’t have the resources to employ a consultant – or even run the tech – required to do it.

The genesis of DataDock

So after ODCamp 2 in Manchester, Jen said she had gone away to try and make things easier, and to build something that people could use. This tool, DataDock, is now available at datadock.io. And it addresses the cost issue by putting data into users’ own GitHubs.

“It leads on from the last session [a debate on whether open data should be published through portals, which concluded that while portals might be useful for some users, who wanted technical specifications and data sets, it was less likely to help less tech-savvy users who wanted a bit of data for a campaign, report or article].

“It lets people publish their own data, instead of getting them to put it all in once place, because the standards are out there to do this… we just have to make it accessible to them.”

Linked suggestions for linked data

Having got that far, she added, she was keen to see what participants might do with this kind of approach. Participants suggested it would be useful to be able to find out whether somebody had solved a problem they were looking to solve; such as define the boundaries of a city.

Others suggested it would be useful to add a tool that suggested other data sources or parameters to users, depending on their behaviour. “A bit like the old Microsoft Clippy tool… it looks like you’re trying to this, can we help with that?” Jen joked.

Naturally, it’s difficult for users to define what they’d like from a new product, or to predict exactly what they could do with it. The new product, though, is out in alpha, it’s free, and its developers are keen to get feedback. So the message was have a look, come up with ideas, think about how it could be made sustainable, and be ready for the next OD Camp…