I learned how to draw a network diagram with gephi. Today I learn to make a complex point. At first glance, it makes people feel that it is an “upper class” network diagram of a high-level article. I learn to draw the following Cherry Blossom diagram.
The following is the edge file of the original data of this figure:
Operate as in the previous article. Import the edge file. You can see a total of 1048 nodes and 1672 edges.
Then calculate some statistical parameters, such as average degree, modularity, etc.
Then, adjust the color and size of the points. Under the partition option, select the calculated modularity data to color the network diagram by module.
Under the ranking option, map the size of the point to the degree, and manually adjust the maximum and minimum size of the node:
Then, adjust the layout of the network diagram. Here, change the layout to Yifan Hu, and keep the default parameters. Is there such a big paper meaning. (Note:What I can’t grasp well is what kind of data type these layouts are suitable for and what kind of data nature they are more suitable for）
For example, this is the effect of selecting forceatlas 2:
This is the effect of fruchterman Reingold:
It may be better to look at your own data set and the problems you want to explain, and emphasize the parts you want to emphasize most to explain your results.
Then fine tune it by myself. For example, I moved some miscellaneous edges, which looked less rough.
Finally, import the vector diagram:
This kind of layout seems to be more suitable. There are few hub nodes. Most networks belong to a subset of these hub nodes, which is a bit like a multidimensional Venn graph.
This article usesArticle synchronization assistantsynchronization