How to plot one factor level as a base on geom_col/geom_area

Question

I've written following function that makes a customised stacked plot:

stacked_plot <- function(data, what, by = NULL, date_col = date, date_unit = NULL, type = 'area'){

  by <- enquo(by)
  what <- ensym(what)
  date_col <- ensym(date_col)
  date_unit <- enquo(date_unit)
  
  if (!rlang::as_string(date_col) %in% names(data)){
    return(cat('Nie odnaleziono kolumny "', as_string(date_col), '".', sep = ''))
  }
    
  if (!rlang::quo_is_null(date_unit)){
    data <- data %>%
      mutate(!!date_col := floor_date(!!date_col, unit = !!date_unit, week_start = 1))
  }
  
  if (!rlang::quo_is_null(by)) {
    data <- data %>%
      filter(!is.na(!!by)) %>%
      group_by(!!date_col, !!by) %>%
      summarise(!!what := sum(!!what, na.rm = TRUE)) %>%
      ungroup() %>% 
      complete(!!date_col, !!by, fill = rlang::list2(!!what := 0))
  } else {
    data <- data %>%
      group_by(!!date_col) %>% 
      summarise(!!what := sum(!!what, na.rm = TRUE)) %>% 
      complete(!!date_col, fill = rlang::list2(!!what := 0))
  }
  
  if (type == 'area'){
    p <- data %>%
      ggplot(aes(!!date_col, !!what, fill = !!by)) +
      geom_area(position = 'stack')
  } else if (type == 'col'){
    p <- data %>%
      ggplot(aes(!!date_col, !!what, fill = !!by)) +
      geom_col(position = 'stack')
  }
  
  p <- p +
    scale_x_date(breaks = '1 month', date_labels = '%Y-%m', expand = c(.01, .01)) +
    theme_minimal() +
    theme(axis.text.x = element_text(angle = 90, vjust = .4)) +
    labs(fill = '')
  
  return(p)
}

Now, I want to use it with the data like below:

data <- structure(list(category1 = structure(c(7L, 7L, 7L, 7L, 7L, 7L, 
7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 2L, 1L, 8L, 1L, 1L, 
1L, 1L, 6L, 6L, 5L, 5L, 1L, 1L, 8L, 3L, 1L, 1L, 8L, 1L, 1L, 1L, 
1L, 1L, 1L, 4L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 
7L, 7L, 7L, 7L, 7L, 7L, 2L, 1L, 8L, 1L, 1L, 1L, 1L, 6L, 6L, 5L, 
5L, 1L, 1L, 8L, 3L, 1L, 1L, 8L, 1L, 1L, 1L, 1L, 1L, 1L, 4L, 7L, 
7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 
7L, 2L, 1L, 8L, 1L, 1L, 1L, 1L, 6L, 6L, 5L, 5L, 1L, 1L, 8L, 3L, 
1L, 1L, 8L, 1L, 1L, 1L, 1L, 1L, 1L, 4L, 7L, 7L, 7L, 7L, 7L, 7L, 
7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 2L, 1L, 8L, 1L, 
1L, 1L, 1L, 6L, 6L, 5L, 5L, 1L, 1L, 8L, 3L, 1L, 1L, 8L, 1L, 1L, 
1L, 1L, 1L, 1L, 4L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 
7L, 7L, 7L, 7L, 7L, 7L, 7L, 2L, 1L, 8L, 1L, 1L, 1L, 1L, 6L, 6L, 
5L, 5L, 1L), .Label = c("base", "cat1", "cat2", "cat3", "cat4", 
"cat5", "cat6", "cat7"), class = "factor"), date = structure(c(14403, 
14403, 14403, 14403, 14403, 14403, 14403, 14403, 14403, 14403, 
14403, 14403, 14403, 14403, 14403, 14403, 14403, 14403, 14403, 
14403, 14403, 14403, 14403, 14403, 14403, 14403, 14403, 14403, 
14403, 14403, 14403, 14403, 14403, 14410, 14410, 14410, 14410, 
14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 
14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 
14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 
14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 14410, 
14410, 14410, 14410, 14417, 14417, 14417, 14417, 14417, 14417, 
14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 
14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 
14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 
14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 14417, 
14417, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 
14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 
14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 
14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 
14424, 14424, 14424, 14424, 14424, 14424, 14424, 14424, 14431, 
14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 
14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 
14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 
14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 14431, 
14431, 14431), class = "Date"), value = c(0.0296166578938365, 
7.02892806393191e-05, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, -23.1966033032737, 0, -17195.0853457778, 0, 0, 0, 0, 0, 
7861.28404641463, 12189.6349251651, 0, 0, -3741.93702617252, 
0, 176.303827249194, 391.710849761278, 131970.980379196, -1587.22123177257, 
297.978554303167, -51860.1739251141, 0, 0, 0, 0, -391.332709445819, 
0.000172964963558834, 0.0098722192979455, 2.34186560613466e-05, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -7.73219962306076, 
0, -17218.0930016352, 0, 0, 0, 0, 0, 7781.23968988082, 12189.6349251651, 
0, 0, 0, 0, 449.478850296707, 293.783137320959, 131970.980379196, 
-1404.7589064091, 250.836431075847, -56540.9156671359, 0, 0, 
0, 0, -558.95740304599, 5.77335368827169e-05, 0.00329073976598183, 
7.79511453535577e-06, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, -2.57739987435359, 0, -17241.1006574926, 0, 0, 0, 0, 0, 
6598.97373566299, 12189.6349251651, 0, -3324.25546024928, 0, 
0, 549.603379062553, 195.855424880639, 131970.980379196, -529.148187957385, 
219.828510450391, -64437.2982346174, 0, 0, 0, 0, -1447.22409849783, 
1.92288024882845e-05, 0.00109691325532728, 2.60503400284112e-06, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -0.859131813420729, 
0, -17264.10831335, 0, 0, 0, 0, 0, 5437.37054226604, 0, 0, 0, 
0, 0, 293.381058210822, 293.783137320959, 131970.980379196, 526.728756878514, 
207.979955414647, -65107.9475533677, 0, 0, 0, 0, -336.514645781955, 
6.40960082942816e-06, 0.000366094798965479, 8.69455082789682e-07, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -127.057071107617, 
0, -17287.1159692073, 0, 0, 0, 0, 0, 5343.46624155083, 0, 0, 
0)), class = "data.frame", row.names = c(NA, -201L))

so I make the following plot:

data %>% stacked_plot(value, category1, date, type = 'col')

And here's my problem. I can't figure out in what order my factor variable (category1) is stacked. And what I would like to do is to reorder the factor levels within my function so that base category would be always displayed as starting from 0 and the rest of levels would be stacked on or below it. Well, it doesn't always have to be named as base, but I think we can add an argument to our function and supply it with the name of base variable. Of course, input data file can have different number of categories.

score 4 · Answer 1 · answered Nov 21 '18 at 15:51

Take a look at the answer by @Inhabitant in this question: How to control ordering of stacked bar chart using identity on ggplot2

Basically the categories are stacked based on the order of the levels in the factor, and the stacking order starts from top to bottom.

Here's how I reordered the stacking with your data:

df_0 <- df_0 %>%
  filter(!is.na(category1)) %>% 
  group_by(date, category1) %>%
  summarise(value := sum(value, na.rm = TRUE)) %>%
  ungroup() %>% 
  complete(date, category1, fill = rlang::list2(value := 0))

df_0$category1 <- df_0$category1 %>% 
  factor(levels = c("cat1", "cat2", "cat3", "cat4", "cat5", "cat6", "cat7", "base"))

df_0 %>%
  ggplot(aes(date, value, fill = category1)) +
  geom_col(position = 'stack')

Two remarks:

I changed the name of the data from data to df_0 to avoid confusion with the R function data()
To make it easier for myself I worked with the data immediately without the function, but of-course all can be integrated with the function

How to plot one factor level as a base on geom_col/geom_area

1 Answers1