[Andee Rubin]

I’ve been working with ordered categorical distributions with middle school students – not so much because I’ve wanted to, but because the dataset I chose to work with happens to have a fair number of such variables (e.g. small, medium, large).  Categorical data can be tricky – partly because the categories aren’t necessarily ordered appropriately on the axis (since by default they’re ordered alphabetically).  Is there a way to specify the order of categories for a categorical variable?  (And any other hints for working with ordered categorical variables?)

[Dan Damelin]

Once you have created a graph with a categorical axis. You can drag the categories into a particular order. If you only have a few categories that might be OK. Interesting idea about having a setting for ordering the categorical variable once plotted. What kinds of settings would you imagine?

[Andee Rubin]

Taking a hint from Excel (which I rarely do), I was thinking that one type of attribute could be “ordered” and you could enter in a list the order from smallest to largest of the attribute values.  I would do this at the attribute definition level, not at the graph level, since it’s a pain to do every time you plot an attribute.  So I’d want to enter the list (small, medium, large) just once in the attribute definition.

[Author]

Posts