I'm trying to use neovis to visualize a small set of sample data (27k rows) from system tables of a warehouse. The goal is to see which tables from which DBs interact the most with each other (data transactions using scripts).
To that end, the graph pattern is (Object 1)-[:INTERACTS_WITH]->(Object 2). An object in this case is simply the concatenation of DB name and the table name.
I ran pagerank and community detection algorithms and used one of the form examples from the neovis github page to visualize the results. The visual is way too cluttered. as you can see below -
I'm thinking that there are too many communities going on in there? Also out of 5000 objects on each side, the max times one object has ointeracted with another object is 12. I guess this is why the edge (relationship line) thickness between different nodes is barely distinguishable in the image? Should I go back to my data model? Tips?
Thank you very much!