Uniform data tags needed by authorities
The Greater London Authority’s initiative to make more data about public services available to citizens is great news (Boris Johnson opens up public databases), but just having the data available is not enough.
With information being supplied by different councils, there is a danger that the tags used to identify each piece of information are not consistent, making the provision of accurate search results very difficult.
The use of common identifiers will make it easier to find relevant information by bringing together all data about a particular topic.
However, to make all of this knowledge as useful and useable as possible, information about each identifier needs to be published and aggregated. Facts, figures and ideas can only be linked properly if everyone knows the right identifiers to use. For example, if a citizen conducts a search for “rubbish collection”, the site will provide a list of all pages that match the identifier for rubbish, but if another council uses a different tag for related services such as recycling, this information might not be displayed.
While the Greater London Authority is making all of this information publily available, it needs to ensure that each piece of data is using a common, agreed identifier for core subjects. It is only by being able to correctly use these identifiers that the full potential of this data will be unlocked.
Kal Ahmed
Whilst it is vital to use the same identifiers, this can only be achieved if everyone knows what particular identifier everyone else is using. What could put a spanner in the works is if some sites don’t play ball and decide to not publish their identifier data. Whilst the BBC may be making use of DBpedia to publish its identifiers, the full value of these can only be realised if those websites with massive amounts of data, like data.gov.uk, also publish their identifiers in a free and open way. The linked data concept won’t get off the ground unless this happens. Those websites that don’t choose to embrace the free and open nature of linked data and identifiers will get left behind as more and more sites choose to publish their identifier data.
Graham Moore,
Director,
Networked Planet (http://www.networkedplanet.com)
Posted by: Graham Moore | Thursday, 28 January 2010 at 04:53 PM
Are we nearly there yet?
I'm a member of the team that has been working on the Greater London Authority Datastore (data.london.gov.uk). I'm responding to Kal's posting about the need for uniform data tags in our data.
The short answer is: yes, we agree. Throughout the project we've tried to make sure we involved with the developer community who will be using our data. We hosted a well attended open morning in October that was advertised on Twitter. A key message from the developers at that meeting was to "go ugly early". That's just what we've done with over 200 datasets that will be available at the official launch on 8 February. These are mostly in cvs format but we are continuing to add xml documents. We know there is more work to be done on the structure of the xml data.
We've identified a smaller number of key GLA datasets that we believe are unique or of special interest. They are already available but we plan to invest more time on these datasets and also make them available as linked data. That will take a little more time because we want to be sure that we really add value. This is the "make pretty later" part of the project.
The Mayor has made a clear commitment to open data with the intention that the Datastore will also carry datasets from other organisations in the Greater London Authority group, including Transport for London. We hope that by going early with our data we will encourage other organisations to start on the journey towards open data.
So in answer to the question: no, we're not nearly there yet! It's a long journey but we're on the way.
There's also a feedback page and Google group available through the new site where we would be happy to carrry on this discussion.
Graham Lane,
GLA IT Unit
Posted by: Graham Lane | Tuesday, 02 February 2010 at 12:43 PM