sugeerth / State-based-Networks-Analysis

A simple tool to visualize large sclae time-varying networks for advanced analysis

Difference Visualization With a Novel Context Analysis Graph for Interconnect Data

Please click to watch the overview video

DyNet

The advent of recording technologies has given rise to large amounts of rich temporal graph-structured data in many applications. The analysis of such complex networks is challenging due to its complex time-varying behavior. In an internet graph, for example, communications patterns between computers frequently change, hubs or routers change roles transtitioning the entire topology of the entire network. To clearly understand the effect of such temporal changes, we need methods that not only identify changes but also reveal the topological effect the changes have on the network. In this paper, we present a novel difference graph approach for network analysis, that visualizes and reveals the topological effect that a change brings about in a network. We introduce an umbrella metaphor to determine how low-level changes in the graph change the community structure of the graph. We further introduce interaction and visual design techniques to reveal the holistic changes between subsequent graphs. Our method is complemented by interactive techniques that reveal holistic changes describing the cause and effect of topological properties of graphs. We apply our algorithm to two datasets, an internet dataset and a network data set.

Required dependencies

  numpy
  networkx 
  community

Getting Started

Note: Tested on OS X 10.11.6 and Ubuntu 14.04

  pip install flask
  pip install numpy
  pip install networkx
  pip install community
  pip install bctpy
  pip install colorBrewer
  pip install networkx
  python main.py

#Download TimeSum repository and then goto src folder 

DyNET/src> main.py 
	Happy Analysis! 

Contributing

See Contributing

Citation Information###

Please cite TimeSum in your publications if it helps your research:

The TimeSum Project makes use of the following libraries

• Numpy, Copyright (C) 2004-2016, NetworkX is distributed with the BSD license
• PYTHON LOUVAIN, Copyright (c) 2009, Thomas Aynaud [e-mail]