[R] A plot similar to violin plot

[David Winsemius]

On Mar 8, 2011, at 9:51 PM, Ista Zahn wrote:

> I'm afraid I still don't get it (see in line below).
>
> On Wed, Mar 9, 2011 at 1:49 AM, C.H. <chainsawtiney at gmail.com> wrote:
>> Sorry for being ambiguous.
>>
>> The data are some continuous variable (Y-axis) categorized into 3
>> groups (Controls, Depressed EF and Preserved EF, X-asis).
>
> OK. I'm having trouble way before this though. Even taking just one
> group, I still don't understand what this graphic is showing.
>
>>
>> The band of dots on the plot is the data point.
>
> This is where I'm getting lost. You have a *band* of dots for each
> data point (singular)? My head is exploding thinking about this. Why
> do you have mutiple dots for each point?
>
> The density of dots
>> and the "fatness" of the band present the frequency of a particular
>> value in Y-axis.
>
> But the Y-axis variable is continuous. In order to have counts at
> different values of y you will need to bin it somehow. And even then
> you will still have only *one* value for the count in each bin. So
> again, how and why are there muliple dots in each band (your term) /
> bin (my term)?
>
> This property is similar to the violin plot: showing
>> the probability density of the data at different values. Instead of
>> showing a shape in violin plot, this plot shows the actual
>> distribution of the data points.
>
> OK, so let's say we bin y and get the counts in each bin. We can plot
> a point for each point that goes into that count, but what determines
> the position of each point on the x-axis?
>
> It's entirely possible (even probable) that my poor little brain just
> can't wrap itself around the idea. Maybe someone else can chime in and
> help me out.

I was think that one might use jitter on the x coordinate (which is  
going to be a discrete value and if its a factor then:

  plot(jitter(as.numeric(...))

It's not going to be in neat bands at each pseudo-continuous level. I  
have a fair amount of data that has values rounded to the nearest  
tenth, so my plots might look like the example graphic offered, if  
there were such a function.

(I have avoid taking on this challenge so far because there was no  
example data.)

  y <- round(rnorm(300), digits=1)
  plot(jitter(rep(1,300), 5), y, xlim=c(0,2) )

Or I suppose you could table() the y values, set up a blank plot with  
ylim set to range(y), and then "walk through" the entries with the  
points() function.

-- 
David.

>
> Best,
> Ista
>
>>
>> Thank you.
>>
>> Regards,
>>
>> CH
>>
>>
>> On Tue, Mar 8, 2011 at 11:14 PM, Ista Zahn  
>> <izahn at psych.rochester.edu> wrote:
>>> Hi,
>>> I don't understand what is being plotted here. Can you describe what
>>> you want in more detail?
>>>
>>> Best,
>>> Ista
>>>
>>> On Tue, Mar 8, 2011 at 7:27 AM, C.H. <chainsawtiney at gmail.com>  
>>> wrote:
>>>> Dear R Users,
>>>>
>>>> I would like to know is there any package to create a plot like  
>>>> this?
>>>>
>>>> http://dl.dropbox.com/u/5409929/cs1160521f01.gif
>>>>
>>>> X axis is categorical. And the positions of the points are
>>>> corresponding to the frequency. (similar to violinplot)
>>>>
>>>> Thank you.
>>>>
>>>> Regards,
>>>>
>>>> CH
>>>>
>>>> --
>>>> CH Chan
>>>>
>>>> ______________________________________________
>>>> R-help at r-project.org mailing list
>>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>>> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
>>>> and provide commented, minimal, self-contained, reproducible code.
>>>>
>>>
>>>
>>>
>>> --
>>> Ista Zahn
>>> Graduate student
>>> University of Rochester
>>> Department of Clinical and Social Psychology
>>> http://yourpsyche.org
>>>
>>
>>
>>
>> --
>> CH Chan
>>
>
>
>
> -- 
> Ista Zahn
> Graduate student
> University of Rochester
> Department of Clinical and Social Psychology
> http://yourpsyche.org
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

David Winsemius, MD
Heritage Laboratories
West Hartford, CT