Indices of Deprivation - LinkedData demonstrator

[Steve Peters]

Hi All

Thought you would be interested in a new app I've developed, which
blends together the Indices of Deprivation, data (from OpenlyLocal) on
local councillors; data (from theyWorkForYou (on MPs); and data (from
EDUBASE) on schools and other educational establishments.

More info here - http://openviz.wordpress.com/2011/09/08/indices-of-deprivation-linked-data-prototype/

[Kingsley Idehen]

Nice visualization effort!

BTW -- would it be possible for you to make the Linked Data URIs behind
your visualizations available to user agents via at least one of the
following:

1. <link/> entries in <head/> -- using Web Linking pattern to expose
URIs of the linked data sources behind the visualizations

2. "Link:" headers in HTTP response metadata -- this depends on what
control you have over the data server

3. HTTP content negotiation -- ditto

4. Microdata or RDFa based data islands in the HTML docs you are
generating.

Goal is for your resource to deliver good visualization while also
enabling user agents to discover data sources behind presentation pages
and explore them further using linked data's follow-your-nose pattern.

--

Regards,

Kingsley Idehen
President& CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen

[Melvin Carvalho]

On 8 September 2011 17:17, Steve Peters

Great job! One of the best visualizations I've seen to date.

>

[BillRoberts]

+1 to Melvin - I really like what Steve has done.

Quick response to Kingsley:

I agree with your general point that visualizations should always link
to the underlying data to allow deeper investigation by people/
machines - but just to note that there are already quite a few data
links in Steve's visualisation - using the 'standard' 3-blobs triple
icon. Clicking on those takes you to a content-negotiable URI for the
dataset or individual observation.

One thing I particularly like about this app is that it is purely HTML
and Javascript - it doesn't need any server-side stuff to work as it
pulls all its data from public APIs and so is an example of how
powerful visualizations can be spread around the web, embedded in
blogs etc etc.

Regards

Bill

[Kingsley Idehen]

On 9/8/11 1:13 PM, BillRoberts wrote:
> +1 to Melvin - I really like what Steve has done.

Ditto.

> Quick response to Kingsley:
>
> I agree with your general point that visualizations should always link
> to the underlying data to allow deeper investigation by people/
> machines - but just to note that there are already quite a few data
> links in Steve's visualisation - using the 'standard' 3-blobs triple
> icon.

Yes, but that doesn't deliver the fidelity of Web Linking patterns
(which is what <link/> and <head/> are about) for user agents,
especially those that are Linked Data savvy.

> Clicking on those takes you to a content-negotiable URI for the
> dataset or individual observation.

I want a machine agent to figure out and follow its nose courtesy of
relation semantics.

> One thing I particularly like about this app is that it is purely HTML
> and Javascript - it doesn't need any server-side stuff to work as it
> pulls all its data from public APIs and so is an example of how
> powerful visualizations can be spread around the web, embedded in
> blogs etc etc.

Yes, the visualization is great. I am just want it to play well with
Linked Data aware agents using established patterns. Web Linking is
there to be exploited. Every resource should be a conduit to Linked
Data, to the degree possible :-)

Links:

1. http://tools.ietf.org/html/rfc5988 -- Web Linking and associated
Relations .

Kingsley
> Regards
>
> Bill

[Feargal Hogan]

Great stuff Steve.

Fantastic app

Have you ever thought about splitting the IMD into its constituent parts and allowing access to those.

It might help indicate what the major drivers are in particular areas, i.e. is a place particularly unhealthy, or susceptable to crime, or whatever?

That might be helpful

Feargal

[BillRoberts]

Hi Feargal

On Sep 9, 2:16 pm, Feargal Hogan <fear...@thehoganfamily.info> wrote:
> Great stuff Steve.
>
> Fantastic app
>
> Have you ever thought about splitting the IMD into  its constituent parts and allowing access to those.

Not sure if I've understood you correctly, but the various sub-domains
of the IMD (crime, health, employment etc) are all available already
via Steve's app (drop-down list at top left) and also as Linked Data
at http://opendatacommunities.org/datasets

Cheers

Bill

[Feargal Hogan]

Bill
So it does.

I honestly did look for it first time around!

Missed it then. Got it now.

Thanks

Feargal