Thoughts on planned Legend Generator in v1

[Gogolo]

Hi

Unfortunately Im not able to code it, but I have some ideas concerning the planned legend generator in v1.

It would be very useful, especially for using Gephi in publications to have legends. But they not only should be generated but there should be the possibility to configure and edit the legend by hand. Moving legend elements free or in a grid. Writing legend labels and titels with different font, color and size. Adding and removing elements, zooming them. Adding or removing lines between elements. Finally put the legend in Preview on a choosen place, adding a border around the legend. Exporting the network and legend as one pdf or in seperate pdfs.

This is more a legend editor than a generator

This is the wishlist...

edit: I just read the post: http://forum.gephi.org/viewtopic.php?f=31&t=398
which is exactely what I wrote and more... (sorry, in German we say: Eulen nach Athen tragen - what means everything was already said...)


Greetings
Olivier

[mbastian]

Hi Olivier,

I completely agree, would you like to go further and start some specification page on our wiki? The aim would be to make a clear plan and discuss different options with everyone. Then, half of the work is already done and coding could be done for instance for the Google Summer of Code.

[Gogolo]

Hi

If I find time for it. I'm very busy these days...
Best Regards
Olivier

[elijah]

I don't know if this is more feasible or less, but I've found that I can build a legend by grabbing the distinct values of the category of data being displayed and creating nodes for those distinct values. I'm not sure if this is something you could do as an easy, automatic legend, but I like how it looks because the legend itself appears as a series of nodes (Sometimes I put a "Legend" node at the center, so all the legend nodes are attached to it). It's very clean and easy and works well with categorical data. It works, but not so quick and easily, with continuous data, too, because I can sample the continuous data at quantiles or natural breaks to do the same thing.

Just food for thought...