search

networkcube / vistorian

Interactive open source visualization platform for multivariate dynamic networks.
https://networkcube.github.io/vistorian
93 stars 11 forks source link

[VIST] Node tables #37

Closed benjbach closed 8 years ago

benjbach commented 8 years ago

Issue by benjbach Tuesday Jan 26, 2016 at 19:41 GMT Originally opened as https://github.com/nathalieriche/networkcube-dev/issues/17

Vistorian provides three ways to create a pair of node and edge tables: 1) Users provide only an edge table with source and target nodes (at least). From the columns for node and target, vistorian creates a node table and adjusts the edge table. 2) Users provide only a node table, where each row is a node and the columns are relations. E.g. a row is a person, and the columns are: father, mother, etc.. In this case, columns are interpreted as relations. 3) User provides an edge table for each edge and a node table for each node. Note that here, columns in the node table are not interpreted as relations.

In any case, proper node and edge tables for networkcube must be generated. The code is in vistorian/dataview.ts:saveCurrentNetwork(). Creating node and edge tables for 2) works fine and has been tested some times. 1) and 2) have not been tested.

benjbach commented 8 years ago

Comment by nathalieriche Tuesday Apr 19, 2016 at 17:59 GMT

rick could you test this? Depending on bugs in #2/3 we will remove them from the release

benjbach commented 8 years ago

Comment by RickGMSR Tuesday Apr 26, 2016 at 18:13 GMT

I have been testing with the Mode Table “relationcroisic“ and have run into a few issues.

  1. Because of the table size (136 Rows), selecting more than 3 columns with label type “Relation” causes the app to become extremely sluggish to the point that trying to add anything after that takes 2-3 minutes.
  2. Link Selections show as Numerical 1, 2, 3... instead of string values. This causes Adjacent Matrix to only work under specific zoom conditions image
  3. Dynamic Ego Network shows only Link Selections but no useful data image