Interesting discussion wiki on GIThub

[Peter Neubauer]

Hi all,
Ciro Cattuto form the ISI Foundation is looking for feedback on a
neo4j graph layout for timelines, see
https://github.com/ccattuto/neo4j-dynagraph/wiki/Representing-time-dependent-graphs-in-Neo4j

I think this would be very interesting to discuss. Just sayin'.

Cheers,

/peter neubauer

G:  neubauer.peter
S:  peter.neubauer
P:  +46 704 106975
L:   http://www.linkedin.com/in/neubauer
T:   @peterneubauer

If you can write, you can code - @coderdojomalmo
If you can sketch, you can use a graph database - @neo4j

[Duane Nickull]

+1

Was not sure what the mechanism is for Call for feedback. On the Wiki
page itself?

Duane Nickull
***********************************
Consulting and Contracting; Proven Results!
i. Neo4J, Java, LiveCycle ES, Flex, AIR, CQ5 & Mobile
b. http://technoracle.blogspot.com
t. @duanechaos






On 12-05-07 9:57 AM, "Peter Neubauer" <peter.n...@neotechnology.com>
wrote:

[Peter Neubauer]

Yeah,
seems this is GIThub, so you probably just edit the page and add your
comments, or fork it and make a Pull Request :)

Interesting stuff though!

Cheers,

/peter neubauer

G:  neubauer.peter
S:  peter.neubauer
P:  +46 704 106975
L:   http://www.linkedin.com/in/neubauer
T:   @peterneubauer

If you can write, you can code - @coderdojomalmo
If you can sketch, you can use a graph database - @neo4j

[Javier de la Rosa]

On Tue, May 8, 2012 at 8:52 PM, Duane Nickull <du...@uberity.com> wrote:
> Was not sure what the mechanism is for Call for feedback.  On the Wiki
> page itself?

I just added a new issue with my comment and they answer quickly.


--
Javier de la Rosa
http://versae.es

[Ciro Cattuto]

Thanks Javier, and thanks Peter for pointing to our discussion page on github.

We are trying to assess the viability of Neo4j+Cypher for large-scale graph timelines

such as the ones generated by our proximity-sensing platforms

and we are facing some performance issues with Cypher on relatively

small datasets, that we are trying to iron out.

The request for feedback is a way to open up that discussion

and understan how the data model we describe should be improved,

whether we are facing intrinsic performance limitations of Cypher/Neo4j

for data with power-law distributed node connectivity,

and what we can do about that all.

Thanks !

Best, Ciro

[Ciro Cattuto]

[looks like I managed to accidentally delete a previous post of mine to this thread]

Thanks Javier, and thanks Peter for stimulating some discussion on our wiki page.

We are evaluating the viability of Neo4j+Cypher to deal with graph timelines

from wearable proximity sensors, and we would like to think more in general

about representing and querying graph timelines (i.e., time-dependent graphs) in Neo4j.

In experimenting with the data model described in the wiki page,

we faced rather unexpected (for us) performance issues with some Cypher queries,

even on relatively small graphs such as the test data discussed in the wiki.

We are trying to understand whether we should work more on the data model,

on the queries, or whether we're facing intrinsic performance limitations of Neo4j+Cypher

when dealing with highly-connected data.

Thanks in advance for any feedback or insights !

Best, Ciro