Just finished my initial beta visualization of MySQL user data on a web server.
Noteworthy is the fact that I’ve stopped building graphs with force-directed graph (FDG) software and have started using a new mathematical algorithm / transformation I thought of during my sleep last night.
It turns out that FDG layouts just take too long and are too processor intensive for large graphs; and since I’ve come up with a much faster algorithm which takes seconds to process versus an hour, I consider this graph building process a simple, but major break though. Finally, I have shattered the glass ceiling with regard to FDG performance problems by dropping FDG and creating my own processing technique to build graphs.
In the video in this post, we show live web server MySQL session data. The red nodes are bots and other automated processes (except for Google Bots), the yellow nodes are “Goggle Bots”, the blue nodes are guest users, and the greenish nodes are registered users. I will change all of this in future releases.
Unfortunately, the screen capture video quality is not great. The UI data in the upper right hand side of the display shows information about the session, such as IP address, User Agent, and some basic user information about what the user is doing, all in near real time..
I can do a lot more object enrichment and data fusion now that I have solved my graph building performance problem and have a new graph building algorithm which is very fast and also provides more visual information; and is better suited for CSA.