Options

Can I create stacked histograms?

ChrisNelsonChrisNelson Member Posts: 11 Contributor II
edited November 2018 in Help
I have data from multiple sources that I'd like to display on one histogram but use color to indicate the source.  If source A has data for 1 and 5 and 6 and source B has data for 1 and 3 and 4, and C has data for 5 and 10, I'd like to see 1 have a bar for A in one color and a bar for B in another color, 3 and 4 have bars for B, 5 have a bar for A and C, and 10 have a bar for C.  I'd like them stacked so that the overall height is the total count but I can see which sources made up each bin.  The Histogram Color seems like it might do what I want but it looks weird to me.  I see overlapping bars of varying width rather than stacked bars.  Am I missing some option?

Answers

  • Options
    MariusHelfMariusHelf RapidMiner Certified Expert, Member Posts: 1,869 Unicorn
    Please have a look at the advanved charts for that. If you can't get it by experimenting on your own, please find a comprehensive user manual in the download section of our website.
  • Options
    ChrisNelsonChrisNelson Member Posts: 11 Contributor II
    I've succeeded in creating a chart with the bars for a value or bin colored by another dimension but they are next to each other,not stacked.
  • Options
    MariusHelfMariusHelf RapidMiner Certified Expert, Member Posts: 1,869 Unicorn
    If you have come so far, making the bars stacked is easy. Just go to the series configuration (where you also chose to use bars at all), and under format change "stacking" to "absoute".
  • Options
    ChrisNelsonChrisNelson Member Posts: 11 Contributor II
    Beautiful.

    Why has Grouping reduced to just "Distinct Values"?  I can no longer see "equidistant fixed bin count"
  • Options
    MariusHelfMariusHelf RapidMiner Certified Expert, Member Posts: 1,869 Unicorn
    Probably because you put a nominal attribute onto that dimension. Since nominal values don't induce any ordering and distances, there is no equidistant binning :)
Sign In or Register to comment.