Re: HCLS - visualization task force?

>> As some of you know, we're already heading down this path in the C-BRASS
"personalized medicine" project (which ends in 2 weeks), as well as with
some visualization tools for SADI/SHARE.  Hopefully we'll have something to
show everyone in a few days/weeks!

Those are exciting news, Mark!
I tend to agree with the dual-approach

SPARQL gives us a table-like view of the data that can be easily plugged
into existing visualization techniques (not to mention a neat way of
filtering and ranking the results according to some criteria, e.g. "only
want to see proteins/genes that are mis-regulated in cancer and whose
binding site can bind to drug X and rank them by their toxicity in human
cell lines").

but RDF based visualizations can make obvious links that were not obvious
before looking at the data in a graph-based visualization (granted we do a
significant amount of filtering of our data in order to remove the "noise"
that RDF based datasets too often throw at us).

Best,
Lena


On Wed, Mar 14, 2012 at 1:55 PM, Mark <markw@illuminae.com> wrote:

> On Tue, 13 Mar 2012 13:09:46 -0600, Jan Aerts <jan.aerts@esat.kuleuven.be>
> wrote:
>
>
>  regarding the visualization of networks and RDF data. We hope to tackle
>> the
>> visual exploration of semantic web instance data
>>
>
>
> Hi all!  Just wanted to toss my hat in this ring also, because the way Jan
> described their approach caught my attention :-)
>
> My feeling is that attempts to visualize RDF networks isn't going to get
> us very far - the granularity of RDF data (when modeled
> semantically-transparently) is simply too fine and at too "low" a level for
> useful human interpretation.  It seems to me that the way to go is to
> attach visualization tools to a higher level of interpretation, where there
> is some "meaning" to be gleaned from chunks of RDF.  e.g. attach a renderer
> to a particular OWL class, and then all of the *instance data* (as Jan
> says) for that class can be explored with some sense of context.
>  Alternately, you might attach renderers to SPARQL queries in much the same
> way.
>
> As some of you know, we're already heading down this path in the C-BRASS
> "personalized medicine" project (which ends in 2 weeks), as well as with
> some visualization tools for SADI/SHARE.  Hopefully we'll have something to
> show everyone in a few days/weeks!
>
> Mark
>
>


-- 
Helena F. Deus
Post-Doctoral Researcher at DERI/NUIG
http://lenadeus.info/