R: mapped_discrete` objects can only be created from numeric vectors - r

Question

I have the following data in R:
df <- structure(list(t0 = c(3.82, -4.88, NA, -3.83, -3.99, NA, NA,
NA, 6.35, 2.47, 0.28, 0.3, NA, 8.31, NA, NA, NA, 2.76, NA, 1.38
), t1 = c(NA, NA, NA, NA, NA, NA, -1.23, 2.19, 4.13, 3.49, -0.42,
NA, 3.78, 2.7, 1.17, NA, NA, NA, NA, NA), t2 = c(-1.85, NA, 1.46,
0.17, NA, NA, -2.81, 1.75, NA, 2.32, -3.08, -1.39, NA, 7.53,
1.77, NA, 0.1, NA, NA, -2.61), t3 = c(-2.05, 3.73, -2.04, -0.22,
-4.29, NA, NA, -0.11, 0.43, NA, -0.78, 3.24, NA, NA, -1.13, 1.09,
NA, NA, 2.7, NA), t4 = c(1.01, -2.77, NA, -3.05, -2.33, 3.78,
NA, NA, NA, NA, -2.04, -4.01, -2.32, 4, -0.28, NA, NA, 9.04,
NA, -4.12), t5 = c(1.56, NA, 4.89, NA, NA, NA, NA, NA, 0.88,
3.15, NA, NA, 2.59, NA, 2.04, NA, NA, NA, -0.26, NA), t6 = c(0.34,
-0.99, NA, 1.93, NA, NA, NA, NA, 0.35, NA, -6.46, NA, NA, NA,
2.57, NA, NA, 4.89, NA, -5.63), t7 = c(0.52, NA, 0.5, 1.85, -6.23,
NA, NA, 1.59, 7.82, 0.82, NA, NA, -1.77, NA, NA, NA, 2.01, NA,
0.7, -1.55), t8 = c(NA, NA, 4.9, -3.93, -8.13, 3.14, 0.03, 1.67,
3.55, NA, -1.55, 2.57, -0.87, NA, 0.71, -0.1, NA, NA, 2.04, NA
), t9 = c(-1.09, NA, -0.52, NA, NA, NA, NA, NA, NA, 2.05, -5.21,
-0.89, -0.03, NA, 0.66, 3.72, -1.96, NA, NA, NA)), row.names = c(NA,
20L), class = "data.frame")
Using the following tutorial (https://jenslaufer.com/data/analysis/visualize_missing_values_with_ggplot.html), I am trying to make a visualization that shows the percentage of missing data:
library(dplyr)
library(ggplot2)
library(tidyverse)
row.plot <- df %>%
mutate(id = row_number()) %>%
gather(-id, key = "key", value = "val") %>%
mutate(isna = is.na(val)) %>%
ggplot(aes(key, id, fill = isna)) +
geom_raster(alpha=0.8) +
scale_fill_manual(name = "",
values = c('steelblue', 'tomato3'),
labels = c("Present", "Missing")) +
scale_x_discrete(limits = levels) +
labs(x = "Variable",
y = "Row Number", title = "Missing values in rows") +
coord_flip()
When I try to see the results, this is the error that I get:
row.plot
Error in `new_mapped_discrete()`:
! `mapped_discrete` objects can only be created from numeric vectors
Run `rlang::last_error()` to see where the error occurred.
Warning messages:
1: In structure(in_domain, pos = match(in_domain, breaks)) :
Calling 'structure(NULL, *)' is deprecated, as NULL cannot have attributes.
Consider 'structure(list(), *)' instead.
2: In structure(in_domain, pos = match(in_domain, breaks)) :
Calling 'structure(NULL, *)' is deprecated, as NULL cannot have attributes.
Consider 'structure(list(), *)' instead.
3: Removed 200 rows containing missing values (geom_raster).
My Question: Can someone please show me what I am doing wrong and how can I fix this error? In the end, I would like to get this kind of picture:

Answer 1

The error is caused by scale_x_discrete.
You don't need it because in your example id is numeric and doesn't have levels as a factor would:
df %>%
mutate(id = row_number()) %>%
gather(-id, key = "key", value = "val") %>%
mutate(isna = is.na(val)) %>%
ggplot(aes(key, id, fill = isna)) +
geom_raster(alpha=0.8) +
scale_fill_manual(name = "",
values = c('steelblue', 'tomato3'),
labels = c("Present", "Missing")) +
#scale_x_discrete(limits = levels)
labs(x = "Variable",
y = "Row Number", title = "Missing values in rows") +
coord_flip()

Answer 2

It looks like you were wanting to produce this plot for missing data on each row rather than for each variable (though I've provided both here). The main issue is that levels is not provided, so we can create that here, then provide as a factor to scale_x_discrete.
library(tidyverse)
output <- df %>%
mutate(id = row_number()) %>%
pivot_longer(-id, names_to = "key", values_to = "val") %>%
select(-key) %>%
group_by(id) %>%
mutate(isna = is.na(val),
total = n()) %>%
group_by(id, total, isna) %>%
summarise(num.isna = n()) %>%
mutate(pct = num.isna / total * 100)
levels <- output %>% filter(isna == T) %>% arrange(desc(pct)) %>% pull(id)
row.plot <- output %>%
ggplot() +
geom_bar(aes(
x = reorder(id, desc(pct)),
y = pct,
fill = isna
),
stat = 'identity',
alpha = 0.8) +
scale_x_discrete(limits = factor(levels)) +
scale_fill_manual(
name = "",
values = c('steelblue', 'tomato3'),
labels = c("Present", "Missing")
) +
coord_flip() +
labs(title = "Percentage of missing values", x =
'Row Number', y = "% of missing values")
Output
Or if you want to do it by variable, then:
output <- df %>%
pivot_longer(everything(), names_to = "key", values_to = "val") %>%
group_by(key) %>%
mutate(isna = is.na(val),
total = n()) %>%
group_by(key, total, isna) %>%
summarise(num.isna = n()) %>%
mutate(pct = num.isna / total * 100)
levels <- output %>% filter(isna == T) %>% arrange(desc(pct)) %>% pull(key)
row.plot <- output %>%
ggplot() +
geom_bar(aes(
x = reorder(key, desc(pct)),
y = pct,
fill = isna
),
stat = 'identity',
alpha = 0.8) +
scale_x_discrete(limits = levels) +
scale_fill_manual(
name = "",
values = c('steelblue', 'tomato3'),
labels = c("Present", "Missing")
) +
coord_flip() +
labs(title = "Percentage of missing values", x =
'Variable', y = "% of missing values")
Output

Answer 3

When I run the code from your tutorial with your data, there is no error. Maybe you want something like this:
library(tidyverse)
missing.values <- df %>%
gather(key = "key", value = "val") %>%
mutate(isna = is.na(val)) %>%
group_by(key) %>%
mutate(total = n()) %>%
group_by(key, total, isna) %>%
summarise(num.isna = n()) %>%
mutate(pct = num.isna / total * 100)
levels <- (missing.values %>% filter(isna == T) %>% arrange(desc(pct)))$key
percentage.plot <- missing.values %>%
ggplot() +
geom_bar(aes(x = reorder(key, desc(pct)), y = pct, fill=isna), stat = 'identity', alpha=0.8, width = 1) +
scale_x_discrete(limits = levels) +
scale_fill_manual(name = "", values = c('goldenrod3', 'firebrick3'), labels = c("Present", "Missing")) +
coord_flip() +
labs(title = "Percentage of missing values", x = 'Variable', y = "% of missing values") +
theme_bw() +
theme(panel.grid = element_blank(),
panel.border = element_blank())
Output: