New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add a local/global scale button to Ranking #541

Closed
mbastian opened this Issue Mar 4, 2012 · 13 comments

Comments

Projects
None yet
3 participants
@mbastian
Copy link
Member

mbastian commented Mar 4, 2012

Then ranking function currently uses only the visible graph for calculating the bounds. That makes comparaison between views difficult because the scale is modified. To address this use case we should have the ability to use ranking on a global scale. A global scale means that the min/max would be computed on the main view rather than on the visible view.

This issue is discussed in this topic:
http://forum.gephi.org/viewtopic.php?f=30&t=1501

@ghost ghost assigned mbastian Mar 4, 2012

mbastian added a commit to mbastian/gephi that referenced this issue Mar 4, 2012

@mbastian mbastian closed this Mar 4, 2012

@seinecle

This comment has been minimized.

Copy link

seinecle commented Mar 5, 2012

Honestly, that's so very useful. Gephi rocks!

@eduramiba

This comment has been minimized.

Copy link
Member

eduramiba commented Mar 5, 2012

Awesome!

@seinecle

This comment has been minimized.

Copy link

seinecle commented Mar 24, 2012

Hi,
I tested the local / global button but and it seems not to work? In the sense, the ranking does not change when the button is toggled from local to global.
To reproduce:

  1. generate a dynamic network
  2. apply a continuous ranking on node's size, based on degree.
  3. select a small time window in the timeline and move it. Observe the change in node's size.
  4. re-do steps 2 and 3, this time with the "local / global" button toggled. The nodes sizes are still the same across the timeline.
@mbastian

This comment has been minimized.

Copy link
Member

mbastian commented Mar 25, 2012

It works for me. The generated dynamic network might be a bad example because for most of it's period the degree is the same (only the edge weight changes). Try to navigate the timeline near the end when the degree does change. You should observe the size changing when the button is toggled.

@seinecle

This comment has been minimized.

Copy link

seinecle commented Mar 27, 2012

Hi,

I am not sure? Please see the test file and commentary I put on the forum:
https://forum.gephi.org/viewtopic.php?f=23&t=1754&start=0

@mbastian

This comment has been minimized.

Copy link
Member

mbastian commented Mar 27, 2012

Hi, thanks for digging into that. For some reason the file you sent me is not automatically recognized as dynamic network and the timeline doesn't seem to activate properly. Will check what's going on.

@seinecle

This comment has been minimized.

Copy link

seinecle commented Mar 27, 2012

yes. I noticed that as well. It is caused when I do the following in the
gephi dev version:

  • manually create this test file in the data laboratory (manually creating
    the 2 nodes and entering the values for the dynamic attribute)
  • exporting this network as a gexf.
  • to fix that, I just open the file, go to the data lab and manually change
    the "Infinity" value to "2010.0", as they were originally. But then, the
    problems with the dynamic ranking I mention on Github and the forum are
    caused after that - I mean, even with a clean-not-bugged gexf file.

A++

Clement


Clement Levallois
Erasmus University Rotterdam
The Netherlands

pro websitehttp://www.erim.eur.nl/ERIM/People/Person_Details?p_aff_id=4321
/ personal website http://www.clementlevallois.net/

twitter and skype: @seinecle
Discover the NESSHI project: http://www.nesshi.eu

On Tue, Mar 27, 2012 at 5:13 PM, Mathieu Bastian <
reply@reply.github.com

wrote:

Hi, thanks for digging into that. For some reason the file you sent me is
not automatically recognized as dynamic network and the timeline doesn't
seem to activate properly. Will check what's going on.


Reply to this email directly or view it on GitHub:
#541 (comment)

@mbastian

This comment has been minimized.

Copy link
Member

mbastian commented Mar 27, 2012

OK I fixed that. An unrelated bug to detect dynamic graphs when only attribute-based.

For our problem I think I found the problem and updated the way we calculate the min and max (and pushed a patch). However for the moment we don't realy support well attribute-based only dynamic networks for ranking. Because we look at enabled filters only you actually can't enable the local scale on your example. Thinking about fixing that...

@seinecle

This comment has been minimized.

Copy link

seinecle commented Mar 27, 2012

I understand and this will need more time for the local scale.
However I tried on the same test file (bug on dyn attribute fixed, thx!), for the global scale, and it seems not to work either (cursor moves along the timeline while a continuous node-resize-ranking is applied on the attribute: the nodes size is not changing).

@mbastian

This comment has been minimized.

Copy link
Member

mbastian commented Mar 28, 2012

I'm working on a patch

mbastian added a commit to mbastian/gephi that referenced this issue Mar 28, 2012

@mbastian

This comment has been minimized.

Copy link
Member

mbastian commented Mar 28, 2012

Ok two things I fixed:

  • The local/global toggle button can now be activated for attribute-only dynamic graph
  • Auto-apply is now working for attribute-only dynamic graph

The min/max seems to be right now. Note that the min/max bound on the ranking panel won't be updated. That to be fixed later. You can see the value with the text label to verify.

@seinecle

This comment has been minimized.

Copy link

seinecle commented Mar 29, 2012

thx! I'll try it asap in one of my projects.

@seinecle

This comment has been minimized.

Copy link

seinecle commented Mar 30, 2012

ok: it works brilliantly. It is extremely insightful to visualize any dynamic attributes, to be able to see how they scale locally relatively to contemporary values, and then, at a touch of a button, how they compare with the global, historical range of values. Major step, thx!!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment