I generate data I'd like to visualize (with dynamic attributes) using gephi.
The code runs in C++ and writes simulation results in binary format to disk. Up to now, I read these files from python for postprocessing.
There will be some/several thousand nodes and edges (appearing over time), with dynamic attributes over several thousand timesteps.
Per node, I will get abt. 20 double valued attributes, changing over timesteps.
Is this large? not sure...
Questions are:
a)
my box operates on debina linux, 16Gb ram, AMD Radeon HD 8570D, 2048 MB HyperMemory graphics card. Will this work out?
b) I was planning to use some of the dynamic attributes to scale and color the nodes. the rest i'd like to see when hovering or selecting nodes. Is this possible? is there a tutorial for this / where ?
c) the (binary) data will grow and I will have different datasets. is it possible to 'configure' the gephi visualisation and apply it (with little effort) to different datasets? like e.g. in kde's kst plotting tool, where I can easily switch data source files and get identically styled plots?
d) currently I write the binary files from c++, open them in python, and was testing gexf export via pygexf. There is almost no documentation on this. In addition, I'm a bit afraid that writing xml like files for such large datasets will result in huge gexf files.
Is there a documented, efficient way to get my data into gephi (possibly without adding java to c++/python already used for my project) in the working style mentioned in c) above?
Thank you!
AlexanderStatistics:Posted by alexpgp — 24 Sep 2016 12:21
]]>