1

I made a violin plot for the first time. But I am not sure how to interpret it. I have three categories: hard, soft, and neutral and for each of the category I have a corresponding summary statistics (pi). Can someone help understand the violin plot. Basically I expect to see that in hard category, pi would be lowest, for soft category pi would be higher than hard and for neutral pi would be the highest. And I am not sure if I see that here. Can someone help me understand this violin plot. Help would be appreciated.

enter image description here

Peter
  • 13
  • 2
    Do you understand how to read a kernel density plot? – mkt Aug 06 '19 at 13:57
  • Please explain what you hope to get from the violin plot that you wouldn't get from a boxplot. – Dave Aug 06 '19 at 13:59
  • 1
    @Dave Multimodality is one possibility. – mkt Aug 06 '19 at 14:02
  • @mkt I meant Peter. – Dave Aug 06 '19 at 14:17
  • I just wanted to see how values are distributed in all of these three categories. And I figured violin plot might be the best way to do it? – Peter Aug 07 '19 at 07:21
  • @Dave I get that - my point was Peter is aware of it or not, kernel density plots can show patterns that are not apparent in boxplots. – mkt Aug 07 '19 at 07:22
  • @mkt I want to see how values are distributed in all these categories and so decided to do a violin plot. But I don't know how to interpret these plots. – Peter Aug 07 '19 at 07:35
  • @Peter You haven't answered my question - do you understand how to read kernel density plots? Or histograms? (Histograms might be easier for you to understand, if you haven't tried them) – mkt Aug 07 '19 at 07:36
  • I don't know about kernel density plots. But I'll plot histogram and see if that makes sense. Thank you – Peter Aug 07 '19 at 07:38
  • @Peter Kernel density plots are smoothed histograms. Your interprety them a lot like you would interpret a histogram. – Dave Aug 07 '19 at 10:00
  • @Dave what do you mean smoothed histograms? – Peter Aug 07 '19 at 18:55
  • @Peter I mean a kernel density plot. It's kind of like a histogram except smooth. Try the following R commands to compare them: set.seed(2019); x <- rchisq(10000,7); hist(x,main="Histogram"); plot(density(x),main="Kernel Density Plot"); abline(h=0). – Dave Aug 07 '19 at 21:15
  • Thank you @Dave. It makes a whole lot of sense to me now – Peter Aug 08 '19 at 07:15

0 Answers0