Given two social network graphs (passed as arguments in the GraphML format), cv2sng visually and quantitatively compares them. If you use diff to compare two text documents, then use cv2sng to compare two network graphs.
It is hard to compare two social networks. Using rich visualization techniques at side of know statistical reports can help with the problem.
A world where:
- Comparing the social structure of distinct groups is easy.
- Comparing the social structure of a cohesive group at two points in time is easy.
This Python script was developed to compare two social network graphs, which are loaded from GraphML files. The comparison focuses on both structural differences (nodes and edges) and attribute differences between the graphs.
vc2sgn is pure Python code built on top of the NetworkX Python library for the creation, manipulation, and study of the structure, dynamics, and functions of complex networks. It imports GraphML files using the read_graphml() function from NetworkX, therefore it vc2sgn can import and compare the following types of graphs:
-
Undirected graphs: Undirected graphs are graphs in which the edges between nodes are not directed. Using NetworkX Graph class for undirected graphs.
-
Directed graphs: Directed graphs are graphs in which the edges between nodes are directed. Using NetworkX DiGraph class for directed graphs.
-
Multigraphs: Multigraphs are graphs that can have multiple edges between the same pair of nodes. Using NetworkX MultiDiGraph class for directed multigraphs.
-
Weighted graphs: Weighted graphs are graphs in which the edges between nodes have weights associated with them. NetworkX supports weighted graphs by allowing the edges to have a weight attribute.
-
Bipartite graphs: Bipartite graphs are graphs in which the nodes can be divided into two disjoint sets, such that every edge connects a node in one set to a node in the other set. When loading those graphs, the NetworkX bipartite module is uded for modelling them.
vc2sng should work with networks of organizations, networks of individuals and networks of things. But one level at the time.
vc2sng was initially designed to compare social networks of software developers modelled as two-mode networks mined from Git Repositories with ScrapLogGit2Net (see https://github.com/jaateixeira/ScrapLogGit2Net). ScrapLogGit2Net accepts git logs as input and provides GraphML files as output. The output of ScrapLogGit2Net captures who works with who (i.e., who co-edits source-code files with who) in a Git repository. Its outputs as GraphML files are ready to use with vc2sng.
vc2sng outputs 4 plots/visualizations and produces one quantitative report.
- The original imported Graph
- Comparison Graph. All new/added nodes and edges appear with a yellow colour marking and all the deleted/removed nodes appear with red marking. All with adjustable transparency.
- Deleted Nodes/Edges Graph: Contains all nodes and edges that were present in the first graph but were deleted (i.e., do not appear in the second graph).
- Added Nodes/Edges Graph: Contains all nodes and edges that were not in the first graph but were added to the second graph.
- Statistics on the two social network graphs and log information.
This repository contains a Python script called vc2sng for comparing two GraphML networks. The script provides visualizations and detailed comparisons of nodes, edges, and attributes between the two networks. It supports both directed and undirected graphs, as well as weighted and unweighted edges.
- Graph Comparison: Compare two GraphML files and visualize the differences.
- Deleted Nodes/Edges: Generate a network of nodes and edges that were present in the first graph but deleted in the second.
- Added Nodes/Edges: Generate a network of nodes and edges that were added to the second graph but not present in the first.
- Supports Directed/Undirected Graphs: Handles both directed and undirected graphs.
- Supports Weighted/Unweighted Edges: Handles graphs with weighted or unweighted edges.
Ensure you have Python 3.7 or later installed on your system. You can download Python from the official Python website.
pip install networkx matplotlib rich logurugit clone https://github.com/jaateixeira/vc2sng/
cd vc2sng/This Python script is designed to compare two NetworkX graphs, which are loaded from GraphML files. The comparison focuses on both structural differences (nodes and edges) and attribute differences between the graphs. The script provides visualizations, stastistical reports and logs results.
Once the dependencies are installed, you can run the script as follows:
python vc2sng.py <graph1.graphml> <graph2.graphml>python vc2sng.py network1.graphml network2.graphml
This will compare the two provided GraphML files and output the differences, including visualizations of added and deleted nodes and edges. Statistics and execution logs will be also provided on stdout.
- Minjeong Shin, Dongwoo Kim, Jae Hee Lee, Umanga Bista, and Lexing Xie. "Visualizing Graph Differences from Social Media Streams." In Proceedings of the Twelfth ACM International Conference on Web Search and Data Mining (WSDM '19), 806–809. Association for Computing Machinery, New York, NY, USA. 2019. https://doi.org/10.1145/3289600.3290616
- K. Andrews, M. Wohlfahrt, and G. Wurzinger. "Visual Graph Comparison." In Proceedings of the 13th International Conference on Information Visualisation (IV 2009), pp. 62-67. Barcelona, Spain, 2009. https://doi.org/10.1109/IV.2009.108
- Tantardini, M., Ieva, F., Tajoli, L. et al. Comparing methods for comparing networks. Nature Scientific Reports 9, 17557 (2019). https://doi.org/10.1038/s41598-019-53708-y
- Teixeira, J., Robles, G. & González-Barahona, J.M. Lessons learned from applying social network analysis on an industrial Free/Libre/Open Source Software ecosystem. J Internet Serv Appl 6, 14 (2015). https://doi.org/10.1186/s13174-015-0028-2
- Coupette, C., & Vreeken, J. (2021, August). Graph similarity description: How are these graphs similar?. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining (pp. 185-195).
Keywords: Visualization, Prototypes, Design engineering, Maintenance engineering, Tree graphs, Stacking, Information analysis, Business communication, Cognitive science, graph drawing, graph comparison, visual, side-by-side, node equivalence matrix, merge graph, business process models
If you'd like to contribute to this project, please follow the standard GitHub fork and pull request workflow. See contributing.md for the contributing guide to vc2sng
This project is licensed under the GPL-3.0 License - see the LICENSE file for details.