geom_histogram with binwidth function: Example from tidyverse website not working

[emilelatour]

I am trying to run an example that can be found on one of the tidyverse webpages on histograms: http://ggplot2.tidyverse.org/reference/geom_histogram.html

I am trying to use this example to have a function define the binwidth for a histogram. Here is the code example from this link

# You can specify a function for calculating binwidth,
# particularly useful when faceting along variables with
# different ranges
mtlong <- reshape2::melt(mtcars)
#> No id variables; using all as measure variables
ggplot(mtlong, aes(value)) + facet_wrap(~variable, scales = 'free_x') +
  geom_histogram(binwidth = function(x) 2 * IQR(x) / (length(x)^(1/3)))

When I run the last command, I get an empty facetted plot (i.e. no histograms) and warnings(). Here is a snippet from the Console:

There were 11 warnings (use warnings() to see them)
> warnings()
Warning messages:
1: Computation failed in `stat_bin()`:
is.numeric(width) is not TRUE
2: Computation failed in `stat_bin()`:
is.numeric(width) is not TRUE
3: Computation failed in `stat_bin()`:
is.numeric(width) is not TRUE

I raise this as an issue here hoping for a solution, but also to make sure that you are aware of the example also showing on the tidyverse website.

Thanks!
Emile

Session Info:

R version 3.4.2 (2017-09-28)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] OIsurv_0.2      KMsurv_0.1-5    broom_0.4.2     survival_2.41-3 dplyr_0.7.4     purrr_0.2.4     readr_1.1.1    
 [8] tidyr_0.7.2     tibble_1.3.4    ggplot2_2.2.1   tidyverse_1.1.1

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.13       lubridate_1.7.0    lattice_0.20-35    assertthat_0.2.0   rprojroot_1.2     
 [6] digest_0.6.12      psych_1.7.8        R6_2.2.2           cellranger_1.1.0   plyr_1.8.4        
[11] backports_1.1.1    MatrixModels_0.4-1 evaluate_0.10.1    httr_1.3.1         rlang_0.1.2       
[16] lazyeval_0.2.1     readxl_1.0.0       minqa_1.2.4        SparseM_1.77       extrafontdb_1.0   
[21] car_2.1-5          nloptr_1.0.4       Matrix_1.2-11      rmarkdown_1.6      labeling_0.3      
[26] splines_3.4.2      extrafont_0.17     lme4_1.1-14        stringr_1.2.0      foreign_0.8-69    
[31] munsell_0.4.3      compiler_3.4.2     modelr_0.1.1       janitor_0.3.0      pkgconfig_2.0.1   
[36] base64enc_0.1-3    mnormt_1.5-5       mgcv_1.8-22        htmltools_0.3.6    nnet_7.3-12       
[41] MASS_7.3-47        grid_3.4.2         Rttf2pt1_1.3.4     nlme_3.1-131       jsonlite_1.5      
[46] gtable_0.2.0       pacman_0.4.6       magrittr_1.5       scales_0.5.0       stringi_1.1.5     
[51] reshape2_1.4.2     bindrcpp_0.2       xml2_1.1.1         tools_3.4.2        forcats_0.2.0     
[56] glue_1.2.0         hms_0.3            parallel_3.4.2     pbkrtest_0.4-7     yaml_2.1.14       
[61] colorspace_1.4-0   rvest_0.3.2        knitr_1.17         bindr_0.1          haven_1.1.0       
[66] quantreg_5.34     

[ptoche]

I'm on ggplot2 v. 2.2.1 and your code is running fine. Just a little message after mtlong:

library(ggplot2)
mtlong <- reshape2::melt(mtcars)
No id variables; using all as measure variables
#> No id variables; using all as measure variables
ggplot(mtlong, aes(value)) + facet_wrap(~variable, scales = 'free_x') +
    geom_histogram(binwidth = function(x) 2 * IQR(x) / (length(x)^(1/3)))

[emilelatour]

Hmm.... Are you running R on windows or a mac? I've tried it on 3 different windows machines at this point, all running most recent versions of R, RStudio, and ggplot2_2.2.1. I still get

[hadley]

You might need to use the devel version; I'm starting to work on the ggplot2 release now

[emilelatour]

Thank you! That did indeed fix this issue!

When I installed the development version and ran the code again, then things work great! Thank you for such cool features like this!

devtools::install_github("tidyverse/ggplot2")
library(ggplot2)
mtlong <- reshape2::melt(mtcars)
#> No id variables; using all as measure variables
ggplot(mtlong, aes(value)) + facet_wrap(~variable, scales = 'free_x') +
  geom_histogram(binwidth = function(x) 2 * IQR(x) / (length(x)^(1/3)))

[jugi92]

This thread helped me and we can use the nclass functions like:

#install.packages("devtools")
devtools::install_github("tidyverse/ggplot2")
library(ggplot2)

mtlong <- reshape2::melt(mtcars)

ggplot(mtlong, aes(value)) + facet_wrap(~variable, scales = 'free_x') +
  geom_histogram(binwidth = function(x) (max(x)-min(x))/nclass.FD(x))
ggplot(mtlong, aes(value)) + facet_wrap(~variable, scales = 'free_x') +
  geom_histogram(binwidth = function(x) (max(x)-min(x))/nclass.scott(x))
ggplot(mtlong, aes(value)) + facet_wrap(~variable, scales = 'free_x') +
  geom_histogram(binwidth = function(x) (max(x)-min(x))/nclass.Sturges(x))

[emilelatour]

Very cool! Your code is so succinct! I ended up with something similar but I also utilized the pretty() base R function which returns "equally spaced 'round' values..." You can see a very clear difference from your code with the variables drat, wt, and carb. Interesting topic about how the choice of binwidth can affect the appearance of histograms.

# devtools::install_github("tidyverse/ggplot2")
library(ggplot2)

mtlong <- reshape2::melt(mtcars)

ggplot(mtlong, aes(value)) +
  facet_wrap(~variable, scales = "free_x") +
  geom_histogram(stat = "bin", binwidth = function(x) {
    pretty(range(x), n = nclass.FD(x), min.n = 1, right = TRUE)[2] - 
      pretty(range(x), n = nclass.FD(x), min.n = 1, right = TRUE)[1]
  })

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/