GGplot2 facet_wrap() with scaling for each variable

I have a dataframe data with 9 predictor variables and one indicator variable ind with values from 0 to 3. I need to make histograms for all of the variables given a value of ind. I've already wrote some code; having no better idea I just divided my dataset into 4 disjoint subsets.

data_no0 <- data[data$ind == 0, -1]
data_no1 <- data[data$ind == 1, -1]
data_no2 <- data[data$ind == 2, -1]
data_no3 <- data[data$ind == 3, -1]

ggplot(gather(data_no1), aes(value)) + 
  geom_histogram(bins = 6) + 
  facet_wrap(~key, scales = 'free_x')

What's more I want to be able to compare histograms for different values of ind so I would like to fix min and max values of x to be (i.e for the first variable) min(data$avg_f0env_sma0) and max(data$avg_f0env_sma), respectively.

I've tried to do it in that way:

scales_x <- list(
  'avg_jitterlocal_sma' 
  = scale_y_continuous(limits = c(min(data$avg_jitterlocal_sma), max(data$avg_jitterlocal_sma))),
  'avg_jitterddp_sma' 
  = scale_y_continuous(limits = c(min(data$avg_jitterddp_sma), max(data$avg_jitterddp_sma))),
  'avg_shimmerlocal_sma' 
  = scale_y_continuous(limits = c(min(data$avg_shimmerlocal_sma), max(data$avg_shimmerlocal_sma))),
  'avgx_pcm_fftmag_spectralflux_sma' 
  = scale_y_continuous(limits = c(min(data$avg_pcm_fftmag_spectralflux_sma), max(data$avg_pcm_fftmag_spectralflux_sma))),
  'avg_pcm_fftmag_spectralcentroid_sma' 
  = scale_y_continuous(limits = c(min(data$avg_pcm_fftmag_spectralcentroid_sma), max(data$avg_pcm_fftmag_spectralcentroid_sma))),
  'avg_pcm_fftmag_spectralharmonicity_sma_compare' 
  = scale_y_continuous(limits = c(min(data$avg_pcm_fftmag_spectralharmonicity_sma_compare), max(data$avg_pcm_fftmag_spectralharmonicity_sma_compare))),
  'avg_f0final_sma' 
  = scale_y_continuous(limits = c(min(data$avg_f0final_sma), max(data$avg_f0final_sma))),
  'avg_f0env_sma' 
  = scale_y_continuous(limits = c(min(data$avg_f0env_sma), max(data$avg_f0env_sma))),
  'avg_loudness_sma3' 
  = scale_y_continuous(limits = c(min(data$avg_loudness_sma3), max(data$avg_loudness_sma3)))
)

ggplot(gather(data_no0), aes(value)) + 
  geom_histogram(bins = 6) + 
  facet_wrap(~key, scales = list(x = scales_x))

But it doesn't work. I've also read about facet_grid_sc(), but there's also a problem with it.

Looks like your dataset could use some tidying, but data wrangling principles aside, it seems that you are looking to control the scales individually for each facet - am I correct?

The lemon package contains some useful functions including facet_rep_wrap() and facet_rep_grid(). One useful feature of these functions is to allow for each facet to have a differently scaled y or x axis. Here's an example with the diamonds dataset.

Histograms of price according to cut with facet_wrap():

library(ggplot2)
library(lemon)

ggplot(diamonds, aes(x=price, fill=cut)) +
  geom_histogram() +
  facet_wrap(~cut)

And here's the same plot using lemon::facet_rep_wrap() with a free y scale:

ggplot(diamonds, aes(x=price, fill=cut)) +
  geom_histogram() +
  facet_rep_wrap(~cut, scales="free")