[SOLVED] Questions about different layouts

[zamtam]

Hi all,

First of all, congratz for this software,wich is really useful and thoughtful for me as a student in sociology. I started to "play" with gephi two years ago and the new release is really confortable.. However, I'm currently stuck because of the specificty of my datas.

In fact, I'm working with internet user comments that I've qualitatively coded with a CAQDAS software. Then each comments (transformed in node) is linked with differents codes (transformed in nodes too). Then, within the vizualiation, I need that the differents codes are attracting each other to view the different registers used by internet users in theirs comments. The problem is : the nodes figuring these codes are not directly linked but they are linked via other nodes. Then, I would like that 2 codes are attract each other relatively to the amount of nodes that exist between them...

I'm not sure to be clear...But regarding this conception of the datas, I'm wondering wich is the best layout / presets ?

I would be glad to get any advices

thxs

Matthieu

PS : to give an example of my gdf, this is something like that :

nodedef:
comment1
comment2
comment3
comment4
code1
code2
code3
code4
edgedef:
code1,comment1
code1,comment2
code1,comment3
code1,comment4
code2,Comment1
code2,comment2
code2,comment3
code3,comment1

then, I need the code 1 and code 2 are too close each other than with code3 because they share more target....

[Yestin]

Hi,
I suggest having a look at this threadhttp://forum.gephi.org/viewtopic.php?f=4&t=873. The common ground of you is that you both want to display two types of data in one graph.

[Yestin]

Hi,
I suggest having a look at this threadhttp://forum.gephi.org/viewtopic.php?f=4&t=873. The common ground of you is that you both want to display two types of data in one graph.

I think it over again. If you did want to use graph to show such dataset, maybe you can try representing each edge(comment->code) as one node. Though this is not a good idea, but maybe you can find something unusual.

[zamtam]

thanks you for your answer..and for the delay..
I finally transform the amount of "common node" between 2 codes into a edge weight value..and it works. The atlas layout fits perfectly with the wieghted graph.
Thanks again
M