Retain user's ordering of columns in Data Laboratory

[suthers]

Problem: if one uses the Light Bulb icon to uncheck or check columns in Data Laboratory, prior sorting of columns is changed back to default.

When I am examining the results in analysis in Data Laboratory I often drag columns around to put relevant columns next to each other. While teaching I might I compute some centrality metrics and then go into Data Laboratory to compare them, dragging the columns to put metrics next to each other that I want to compare (e.g., weighted to unweighted, directed to undirected, or values from igraph to values from Gephi, or perhaps to show how eigenvector and pagerank differ, etc.). Then if I compute more metrics and use the Light Bulb icon to uncheck some columns so that it is less cluttered, my sorting of columns is wiped out and they are restored to the prior order. I then need to take class time to put things back (taking advantage of the teachable moment to comment on usability issues).

Proposed solution

Retain state that results from user manipulation whenever possible. (A general principle for user adaptive interfaces and to save redoing work that you can also apply elsewhere, e.g., min and max node size defaults.)

Alternatives considered

Additional context