R+ggplot2+facet_wrap: preserve the with of the bars

Question

Please have a look at the reprex at the end of the post. It is a faceted barchart where the bars are placed in decreasing order in each subplot. My only question is how to ensure that the bars have the same consistent width in all the subplots.

I had a look at

How to automatically adjust the width of each facet for facet_wrap?

Equal bar width across facet_wrap barplots

but I did not go very far.

Any suggestion is appreciated!

library(tidyverse)


sec_ms_tcf_plot <- structure(list(member_state = c("BG", "EE", "EE", "HR", "HR", 
"HR", "HR", "HR", "HR", "HR", "LT", "LT", "LT", "LT", "LT", "LT", 
"LT", "LT", "MT", "MT", "MT", "PL", "PL", "PL", "PL", "PL", "PL", 
"PL", "PL", "PL", "PL", "PL", "PL", "PL", "PL", "PL", "PL", "RO", 
"RO", "RO", "RO", "RO", "RO", "RO", "RO", "RO", "RO", "RO", "RO", 
"RO", "RO", "RO", "RO", "RO", "RO", "SE", "SE", "SK", "SK", "SK", 
"SK", "SK", "SK", "SK", "SK", "SK", "SK", "SK", "SK", "SK", "SK", 
"SK", "SK", "SK", "SK"), sector = c("a_agriculture_forestry_and_fishing", 
"c_manufacturing", "a_agriculture_forestry_and_fishing", "b_mining_and_quarrying", 
"d_electricity_gas_steam_and_air_conditioning_supply", "a_agriculture_forestry_and_fishing", 
"g_wholesale_and_retail_trade_repair_of_motor_vehicles_and_motorcycles", 
"f_construction", "h_transportation_and_storage", "c_manufacturing", 
"i_accommodation_and_food_service_activities", "l_real_estate_activities", 
"h_transportation_and_storage", "e_water_supply_sewerage_waste_management_and_remediation_activities", 
"g_wholesale_and_retail_trade_repair_of_motor_vehicles_and_motorcycles", 
"d_electricity_gas_steam_and_air_conditioning_supply", "a_agriculture_forestry_and_fishing", 
"c_manufacturing", "a_agriculture_forestry_and_fishing", "h_transportation_and_storage", 
"c_manufacturing", "q_human_health_and_social_work_activities", 
"b_mining_and_quarrying", "l_real_estate_activities", "i_accommodation_and_food_service_activities", 
"s_other_service_activities", "e_water_supply_sewerage_waste_management_and_remediation_activities", 
"j_information_and_communication", "k_financial_and_insurance_activities", 
"m_professional_scientific_and_technical_activities", "h_transportation_and_storage", 
"n_administrative_and_support_service_activities", "d_electricity_gas_steam_and_air_conditioning_supply", 
"f_construction", "g_wholesale_and_retail_trade_repair_of_motor_vehicles_and_motorcycles", 
"c_manufacturing", "a_agriculture_forestry_and_fishing", "p_education", 
"k_financial_and_insurance_activities", "o_public_administration_and_defence_compulsory_social_security", 
"r_arts_entertainment_and_recreation", "s_other_service_activities", 
"b_mining_and_quarrying", "d_electricity_gas_steam_and_air_conditioning_supply", 
"q_human_health_and_social_work_activities", "e_water_supply_sewerage_waste_management_and_remediation_activities", 
"m_professional_scientific_and_technical_activities", "n_administrative_and_support_service_activities", 
"j_information_and_communication", "i_accommodation_and_food_service_activities", 
"a_agriculture_forestry_and_fishing", "h_transportation_and_storage", 
"f_construction", "c_manufacturing", "g_wholesale_and_retail_trade_repair_of_motor_vehicles_and_motorcycles", 
"a_agriculture_forestry_and_fishing", "d_electricity_gas_steam_and_air_conditioning_supply", 
"p_education", "n_administrative_and_support_service_activities", 
"s_other_service_activities", "j_information_and_communication", 
"m_professional_scientific_and_technical_activities", "o_public_administration_and_defence_compulsory_social_security", 
"f_construction", "r_arts_entertainment_and_recreation", "e_water_supply_sewerage_waste_management_and_remediation_activities", 
"b_mining_and_quarrying", "d_electricity_gas_steam_and_air_conditioning_supply", 
"h_transportation_and_storage", "q_human_health_and_social_work_activities", 
"i_accommodation_and_food_service_activities", "g_wholesale_and_retail_trade_repair_of_motor_vehicles_and_motorcycles", 
"l_real_estate_activities", "a_agriculture_forestry_and_fishing", 
"c_manufacturing"), sector_exp = c(217483, 2374, 3730, 31, 1217, 
1563, 3637, 5181, 5294, 33715, 8, 93, 271, 750, 857, 3180, 32106, 
43777, 715, 759, 14230, 129, 143, 224, 265, 383, 1016, 1569, 
1763, 2254, 2309, 2507, 5307, 15428, 38174, 38189, 558519, 416, 
673, 2194, 2726, 3381, 6841, 9425, 10731, 16835, 18968, 20845, 
23340, 43572, 170800, 175884, 205434, 272582, 892678, 149629, 
149629, 75, 466, 498, 607, 679, 918, 919, 1014, 1037, 1130, 1521, 
1635, 1639, 1837, 3715, 5275, 11874, 51306), sector_code = c("A", 
"C", "A", "B", "D", "A", "G", "F", "H", "C", "I", "L", "H", "E", 
"G", "D", "A", "C", "A", "H", "C", "Q", "B", "L", "I", "S", "E", 
"J", "K", "M", "H", "N", "D", "F", "G", "C", "A", "P", "K", "O", 
"R", "S", "B", "D", "Q", "E", "M", "N", "J", "I", "A", "H", "F", 
"C", "G", "A", "D", "P", "N", "S", "J", "M", "O", "F", "R", "E", 
"B", "D", "H", "Q", "I", "G", "L", "A", "C")), row.names = c(NA, 
-75L), class = c("tbl_df", "tbl", "data.frame"))
 


## See https://stackoverflow.com/questions/5409776/how-to-order-bars-in-faceted-ggplot2-bar-chart/5414445#5414445


dd2 <- sec_ms_tcf_plot |>
    mutate(category2=factor(paste(member_state, sector_code))) |>
    mutate(category2=reorder(category2, rank(sector_exp)))


gpl <- ggplot(dd2, aes(y=category2, x=sector_exp/1e3)) +
  geom_bar(stat = "identity") +
  facet_wrap(. ~ member_state, scales = "free_y", nrow=3) +
    scale_y_discrete(labels=dd2$sector_code, breaks=dd2$category2,
                     )+
        xlab("Sector Expenditure (Mio \u20ac)")+
    ylab("Sector NACE Code")


ggsave( "expenditure.png", gpl, width=10, height=13)



##Is there any way to have the same bar width in all the subplots?

^{Created on 2023-05-26 with reprex v2.0.2}

Jon Spring · Accepted Answer · 2023-06-03T23:35:10.123

1

This is not great but sort of does it.

It's tricky since ggplot bar widths are scaled in proportion to the number of categories in each facet, and you want those to vary. One hack would be to add blank categories into each facet to augment them so that every facet has as many categories as the most crowded facet.

pad_facets <- function(df, facet_var, cat_var) {
  df %>%
    distinct({{facet_var}}, {{cat_var}}) %>%            # keep 1 row per facet/cat
    count({{facet_var}}) %>%                            # how many cat per facet?
    transmute({{facet_var}}, cat_to_add = max(n) - n) %>% # calc # cat to add
    uncount(cat_to_add) %>%                               # copy each cat x times
    mutate({{cat_var}} := as_factor(row_number())) %>%  # add new factor level
    bind_rows(df)                                         # add on original data
}

dd2 %>%
  pad_facets(member_state, category2) %>%

ggplot(aes(y=category2, x=sector_exp/1e3)) +
  geom_bar(stat = "identity",
           position = position_dodge(preserve = "total")) +
  facet_wrap(. ~ member_state, scales = "free_y", nrow=3) +
  scale_y_discrete(labels=dd2$sector_code, breaks=dd2$category2,
  )+
  xlab("Sector Expenditure (Mio \u20ac)")+
  ylab("Sector NACE Code")

edited Jun 03 '23 at 23:35

answered May 26 '23 at 19:47

Jon Spring

55,165
4
35
53

The final result is fairly acceptable. I just wonder if you could explain a bit what you do and above all: is there a way to put everything into a function or something I can reuse? It is a pain to write this for every new panel plot I want to make. – larry77 May 26 '23 at 19:55
Edited to put the padding into a function. I obviously can't guarantee this will work for all cases, but it works for the example data. – Jon Spring May 26 '23 at 20:07
I just do not understand why you write "{{cat_var}}". The function seems to work fine also with {{cat_var}} without the " ". Can you explain why you need the " " ? – larry77 Jun 03 '23 at 13:53
I don't understand what you are referring to. – Jon Spring Jun 03 '23 at 18:03
In your function pad_facets you have the line mutate("{{cat_var}}" := as_factor(row_number())) , but it seems to me that mutate({{cat_var}} := as_factor(row_number())) also works. – larry77 Jun 03 '23 at 20:44
I was going by the example here, but yes it looks like the quotes are optional. https://dplyr.tidyverse.org/articles/programming.html#name-injection – Jon Spring Jun 03 '23 at 23:36

R+ggplot2+facet_wrap: preserve the with of the bars

1 Answers1