Skip to content

GISRUK Linked Data Workshop – some thoughts and conjecture

April 28, 2011

On Tuesday I attended the one day, pre-conference workshop that traditionally accompanies GISRUK. Held in a gloriously sunny Portsmouth, this years workshop focused on the exponentially trending topic of linked data – specifically the potential of geographic data in this context.

The standard of presentations was excellent – lots of food for thought. I won’t go into great detail (Nick Gould from Manchester provides a nice summary of the talks here), but rather just briefly discuss potential relevance to U•Geo (and perhaps data archives in general).

Linked data clearly has immense potential, and the possibilities hinted at got me thinking about Archive data. To take our national level surveys as an example, I’ll give a quick break-down of a theoretically possible use of linked data. Let’s assume all of the datasets have a spatial variable – let’s say an administrative unit such district or ward.

  • Carefully check the integrity of those variables (via something like Unlock) and match each to a time-stamped administrative unit definition based on documentation/other info.
  • Work out, using geo-referenced linked data on administrative geographies, that this area of space has ‘x’ relationship to this other area of space i.e. establish homology. Thus creating an administrative unit crosswalk – enabling statements like: these districts are equivalent to this county.
  • Given these relative relationships, query the data for say, the UK counties (irrespective of whether these counties were included as a variable in the original survey data files) and call back all variables which contain data on say, smoking habits, for these units.
  • You could then very quickly place whatever interesting data you’d located onto a map using the district boundary definitions and Google Maps API.
  • Perhaps the best part though, is that my query could also trawl other linked data repository’s, e.g. dbpedia.org for the entries associated with those (defined) places , and bring back other characteristics of the districts to overlay. You’ve very quickly got this huge, linked up, multi-source data mashup.

The possibilities are pretty much endless.

On another front: while I had established that historical administrative units are a pain in my own time, it was slightly gratifying to have it confirmed by other practitioners! The good news is, there are lots of people who are enthusiastic about improving their documentation, and who are aiming for open and linked data. Despite this hope, it seems that linked geographical data at the Archive may still be a long way off – without standardised URI indexed unit definitions to reference, any such developments are heavily constrained. Let’s hope for an initiative (perhaps government) in this area sometime soon.

Advertisements
One Comment leave one →
  1. April 29, 2011 4:05 pm

    I think that’s the issue for a lot of people, what can you linked data reference is there for you the link to? It’s important people know the rules of linked data (talked about in this video) http://blip.tv/file/5040871 ) so everyone is linking properly.

    You can link to some Ordnance Survey data http://data.ordnancesurvey.co.uk
    Some people consider OpenStreetMap objects to be linked data (eg http://www.openstreetmap.org/browse/relation/127864) although you can’t rely on the object ids staying the same (anyone could delete the object and create it afresh, that would give it a new id number). Documentation for getting the object details/geography is at http://wiki.openstreetmap.org/wiki/Xapi

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: