I'd take a second to think about what you want to visualize. Do you
really want to see a visualization of all the RDF assertions? It can
get pretty messy pretty quick, which tends to make the visualization
pretty useless in my experience. Or perhaps I haven't run across the
right one. I suspect there is some property (or properties) of your
graph that you want to provide insight into?
For example, I was doing some research on the httpRange-14 issue, and
I wanted to do some analysis of conversations about it. So I pulled
down the www-tag discussion list archives, and converted them to RDF
with a script. After doing some basic querying I got interested in an
extensive conversation that happened around a change proposal that was
made by Jeni Tennison -- and I wanted to see what the threaded
conversation looked like as a graph. I really wanted to get visual
insight into who was responding to who, and where the conversation
dead ended, forked, etc.
So I wrote a little adaptor that did a query of the RDF and generated
a JSON file [1] which included a piece of metadata about each email,
and the connections between them. I did the visualization with D3
which has some nice support for graphs. Each node in the graph is
color coded to the sender, and links to the actual email. I did it
statically because my data wasn't changing, but I guess I could have
generated the JSON dynamically from a SPARQL query, or a series of
SPARQL queries.
http://inkdroid.org/rage14/proposal/
So I guess this is a long way of asking why you want to visualize the graph :-D
//Ed
[1] http://inkdroid.org/rage14/proposal.json
On Wed, May 1, 2013 at 9:24 AM, Ethan Gruber <[log in to unmask]> wrote:
> Hi all,
>
> I have a fair amount of data in a triplestore, and I'd like to experiment
> with different forms of visualization. I have found a few libraries for
> visualizing RDF graphs through Google, but they still seem relatively
> rudimentary. Does anyone on the list have recommendations? I'm looking
> for something that can use SPARQL. I'd like to avoid creating duplicates
> or derivatives of data, like GraphML, unless it is possible to render
> GraphML which has been serialized from SPARQL results on the fly.
>
> Thanks,
> Ethan
