Makes graph from wikipedia links.
Gephi successfully loads 9481 nodes and 25648 edges (highest number tested so far) which seems close to its limit.
This project visualizes connections between Wikipedia links using Python to fetch Wikipedia links and Networkx and Gephi to generate a graph. Depth and ratio variables can be changed to determine how many links are searched. The depth variable specifies how many times a link's children is searched. The ratio variable beween 0 and 1 tells what percentage to randomly sample from all links for each page.
- python (v...)
- beautifulsoup (v...)
- requests (v...)
- networkx (v...)
- gephi (v...)
First choose a Wikipedia topic and depth and ratio. For example, if you wanted to search links from https://en.wikipedia.org/wiki/Discord, modify line 124 in main.py with the string. To modify the depth, go to line 125 and type an integer. To modify the ratio, go to line 101 and insert a decimal between 0 and 1.
First parse wikipedia
python main.py && python convert_networkx.py
convert_networkx.py creates a .gexf file which can be read by Gephi. Networkx supports many different graph file types, but .gexf files are the default.
Next, open the .gexf file in Gephi and edit it until it looks complete.
This graph shows related links (nodes) connected to other links by edges. I started with Earth as the Wikipedia subject using a depth of two and a ratio of 0.015. By looking at the connected edges, I can trace a link's connections to its origin.
[TODO] Allow ratio to be modified from main and pass it as a parameter