[R] Faceted bar plot shows wrong counts (ggplot2)

[Helios de Rosario]

Michael,

Thanks for the pointer to the discussion in the ggplot list. It seems
that the reason of this behaviour of facet_grid() is already known and
being discussed by the developers of ggplot2.

facet_grid() reduces the original data frame with unique() before
applying the stats.  If the data frame has any other column that
prevents duplicated rows, counts are correctly computed.

E.g.

diamonds25 <- droplevels(diamonds[1:25,]) # Keep all columns

# Everything else as before:
base  <-  ggplot(diamonds25,  aes(fill  =  cut))  +
 geom_bar(position  =  "dodge")  +
 opts(legend.position  =  "none")
 base  +  aes(x  =  cut)  +
 facet_grid(.  ~  color)


Helios

>>> El día 12/03/2012 a las 20:59, "R. Michael Weylandt"
<michael.weylandt at gmail.com> escribió:
> You get the "good" behavior with
> 
> base + aes(x = cut) + facet_wrap(~ color, ncol = 5)
> 
> so this seems buggy to me.
> 
> If someone here doesn't step forward with more insight, I'd forward
it
> to the ggplot list to see if one of the developers there can give an
> explanation or possibly make the official call that it's a bug.
> 
> There was another report of a possible bug in facet_grid() today
that
> could be related:
>
https://groups.google.com/group/ggplot2/browse_thread/thread/5213ac35da6b36d

> 4
> 
> Michael
> 
> On Mon, Mar 12, 2012 at 7:16 AM, Helios de Rosario
> <helios.derosario at ibv.upv.es> wrote:
>> I have encountered a problem with faceted bar plots. I have tried
to
>> create something like the example explained in the ggplot2 book (see
pp.
>> 126-128):
>>
>> library(ggplot2)
>> mpg4  <-  subset(mpg,  manufacturer  %in%
>> c("audi",  "volkswagen",  "jeep"))
>> mpg4$manufacturer  <-  as.character(mpg4$manufacturer)
>> mpg4$model  <-  as.character(mpg4$model)
>>
>> base  <-  ggplot(mpg4,  aes(fill  =  model))  +
>> geom_bar(position  =  "dodge")  +
>> opts(legend.position  =  "none")
>> base  +  aes(x  =  model)  +
>> facet_grid(.  ~  manufacturer)
>>
>> That example works fine; the bar heights are just the same as the
>> counts in the table:
>>
>> table(mpg4[,1:2])
>>          model
>> manufacturer a4 a4 quattro a6 quattro grand cherokee 4wd gti jetta
new
>> beetle
>>  audi        7          8          3                  0   0     0
>>    0
>>  jeep        0          0          0                  8   0     0
>>    0
>>  volkswagen  0          0          0                  0   5     9
>>    6
>>          model
>> manufacturer passat
>>  audi            0
>>  jeep            0
>>
>> But in other cases this does not occur. For instance, take a small
>> subset of data(diamonds):
>>
>> diamonds25 <- droplevels(diamonds[1:25,2:3])
>> table(diamonds25)
>>           color
>> cut         E F H I J
>>  Fair      1 0 0 0 0
>>  Good      1 0 0 1 4
>>  Very Good 1 0 3 1 4
>>  Premium   3 1 0 1 0
>>  Ideal     1 0 0 1 2
>>
>> And change the variables mapped in the previous plot:
>>
>> base  <-  ggplot(diamonds25,  aes(fill  =  cut))  +
>> geom_bar(position  =  "dodge")  +
>> opts(legend.position  =  "none")
>> base  +  aes(x  =  cut)  +
>> facet_grid(.  ~  color)
>>
>> I see all bars with height = 1.
>> I have ovserved this problem (wrong bar heights, but not always =
1),
>> in other cases when all counts are very small or zero.
>> What's wrong here?
>>
>> Regards,
>> Helios
>>
>> sessionInfo()
>> R version 2.14.2 (2012-02-29)
>> Platform: i386-pc-mingw32/i386 (32-bit)
>>
>> locale:
>> [1] LC_COLLATE=Spanish_Spain.1252  LC_CTYPE=Spanish_Spain.1252
>> [3] LC_MONETARY=Spanish_Spain.1252 LC_NUMERIC=C
>> [5] LC_TIME=Spanish_Spain.1252
>>
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods  
base
>>
>> other attached packages:
>> [1] ggplot2_0.9.0
>>
>> loaded via a namespace (and not attached):
>>  [1] colorspace_1.1-1   dichromat_1.2-4    digest_0.5.1
>> grid_2.14.2
>>  [5] MASS_7.3-17        memoise_0.1        munsell_0.3
>> plyr_1.7.1
>>  [9] proto_0.3-9.2      RColorBrewer_1.0-5 reshape2_1.2.1
>> scales_0.2.0
>> [13] stringr_0.6
>>
>>
>>
>> INSTITUTO DE BIOMECÁNICA DE VALENCIA
>> Universidad Politécnica de Valencia ● Edificio 9C
>> Camino de Vera s/n ● 46022 VALENCIA (ESPAÑA)
>> Tel. +34 96 387 91 60 ● Fax +34 96 387 91 69
>> www.ibv.org 
>>
>>  Antes de imprimir este e-mail piense bien si es necesario hacerlo.
>> En cumplimiento de la Ley Orgánica 15/1999 reguladora de la
Protección
>> de Datos de Carácter Personal, le informamos de que el presente
mensaje
>> contiene información confidencial, siendo para uso exclusivo del
>> destinatario arriba indicado. En caso de no ser usted el
destinatario
>> del mismo le informamos que su recepción no le autoriza a su
divulgación
>> o reproducción por cualquier medio, debiendo destruirlo de
inmediato,
>> rogándole lo notifique al remitente.
>>
>> ______________________________________________
>> 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.

INSTITUTO DE BIOMECÁNICA DE VALENCIA
Universidad Politécnica de Valencia • Edificio 9C
Camino de Vera s/n • 46022 VALENCIA (ESPAÑA)
Tel. +34 96 387 91 60 • Fax +34 96 387 91 69
www.ibv.org

  Antes de imprimir este e-mail piense bien si es necesario hacerlo.
En cumplimiento de la Ley Orgánica 15/1999 reguladora de la Protección
de Datos de Carácter Personal, le informamos de que el presente mensaje
contiene información confidencial, siendo para uso exclusivo del
destinatario arriba indicado. En caso de no ser usted el destinatario
del mismo le informamos que su recepción no le autoriza a su divulgación
o reproducción por cualquier medio, debiendo destruirlo de inmediato,
rogándole lo notifique al remitente.