arbor layout for larget data

Question

I have too many nodes in my data, and I find out it's really not OK to visualize them by using cytoscape js. Does anyone have similar problem, or there are some solution to my problem, see my graph here.

http://raven.anr.udel.edu/~sunliang/biolayout1/biolayout2.php?fileName=usrID/id_1380304481_42_gene_iterm.txt

Any suggestion and ideas is welcome. Thanks.

Answer 1

Arbor, like all forcedirected layouts, are very nice in that they can often generate acceptable layouts. However, there are many problems with them, including

• performance (speed),
• consistency (i.e. JS does not allow you to set a seed to Math.random() for initial positions),
• quality (it's just a physics simulation, and simulations can be overcomplex or otherwise not provide good analogue node positions),
• and so on.

If you have a large dataset, you should consider using one of the alternative layouts. The breadthfirst layout often generates good results. You also have the option of writing alternative layouts yourself that are tailored to your data. It's a very straightforward process that is outlined in the documentation.

http://cytoscape.github.io/cytoscape.js/#extensions/layouts