Re: [igraph] Visualizing large graphs of BigData

[Tamás Nepusz]

This is probably not a good idea - but not because of igraph. The problem is 
that it is really hard to get out anything meaningful from standard 
nodes-and-edges plots if your graph contains 6 million nodes. Even if you can 
find a meaningful arrangement of the nodes (using the DrL layout algorithm in 
igraph - it’s unlikely that any of the other layout algorithms will give you 
any meaningful layout in a finite time), you would need a very large canvas to 
plot your graph on. For instance, if you set the width and height of your graph 
plot to 2000 x 1500 pixels, that would give you 3 million pixels only - half a 
pixel for every node of your graph, and we did not even count the edges.

I think it is better if you first think about what sort of information you want 
to communicate with your plots and then you try to find an alternative 
visualization method that communicates this information efficiently (assuming 
that you want to create these plots not just for the sake of eye-candy). Hive 
plots may be a good starting point:  

http://www.hiveplot.net/  

--  
T.


On Saturday, 28 December 2013 at 21:17, Ahmed Abdeen Hamed wrote:

> Hello friends,
>  
> How large can igraph visualize large graphs from BigData? I have a about 
> 6Million nodes+ graph. I want to see if this is a good idea before I try it 
> and stick my machine in the deep-fry.  
>  
> Thanks very much!
>  
> -Ahmed  
> _______________________________________________
> igraph-help mailing list
> address@hidden (mailto:address@hidden)
> https://lists.nongnu.org/mailman/listinfo/igraph-help