|
Anyone who has stored the corresponding attributes can color the nodes according to the corresponding “partition” directly in the “Overview” tab on “Design”: Color nodes thematically If you have not done the relevant preparatory work, you can select the relevant nodes directly in the “data lab”, edit them with a right click and color them as desired: coloring of nodes in the gephi data laboratory However, it makes more sense to work with attributes and additional columns. Otherwise you will constantly overwrite your changes. Depending on the result, you can color the edges using the same principle so that certain networks are more visible.
A coloring according to topics can then look like this in the end (by the way, I chose the “Fruchtermann India Car Owner Phone Number List Reingold” layout for distribution for all examples): coloring internal links by topic This does have a certain aesthetic appeal, but the demarcation of the subject areas is only partially apparent: of all the colors, there seem to be two areas that mix strongly, particularly in the lower right edge. It would be better if there was a clear demarcation into thematic silos, which in extreme cases could even look like this: internal linking in silos In both cases you will see a colorful circle on the outside. This is because individual documents are untidy (such as print versions). They are nice and colorful, but they shouldn't actually exist (more on this in my blog post on the topic of crawling control ). Weighted evaluation So far so good.
But there is more! Often it's not just about the topic, but also about the weighting. The classic is the internal PageRank, which can be calculated under “Overview” with two clicks: “Start PageRank”, check the “Use edge weight” box and confirm: Calculate pagerank in gephi But other metrics are also interesting for weighting. For example, you can create a numerical prioritization of the most important landing pages (manually) to highlight them. metric as a “ranking”: gradual coloring of nodes in gephi However, your metric will only appear there if you have stored it as a real numerical value (e.g. as a long or integer data type, not as a string).
|
|