These are the main tables. I want to use a box plot, to plot a selected company against the performance of their given industry Segment. For example, Bean would be in the Lumber industry.
Performance in this case is # of lines per so_id. For each so_id, such as so_id 1, it can appear many times in the data. So I use a count of the so_id, so if 1 appears 13 times, then we call that the # of lines in so_id 1.
Using the box plot wizard, with segment as dimension and aggregator, I use this as the expression count(so_id)/ count(DISTINCT(so_id))
since as said before, I want # of lines per so_id
Then I click median mode, graph, and this comes up.