I suspect that the need for tags explorer to display the network is one of the reasons why browsers struggle with high tweet numbers.
If the network calculation was decoupled from the display it might be possible to produce adjacency matrices for larger networks without causing browsers to crash, which could then be opened in software designed to handle them (e.g. gephi). b
I’m keen to keep a web based solution and I think it’s achievable. The current solution has a number of inefficiencies in particular the visualisation library is a very old version and a lot of processing could be saved. Hoping to have a complete rebuild ready by the end of the year…
… I should add that as well as a better visualisation tool I’m keen to implement data export options for tools like gephi and nodexl
Maybe it’s a D3.JS issue ? another similar project that have similar issue is http://flocker.outliers.es and with this harvesting tool the issue is the same (both Tags & Flocker use D3.js) I was wondering if OGMA from the guys at Linkurious could be a better engine to render big visualizations ?
see : https://linkurio.us/blog/ogma-js-library-large-scale-graph-visualization/
You must be logged in to reply to this topic.