CODEWITHSUNDEEP
rplotmissing-data
AP38
AP38

Reputation: 169

Grid plot of missing data

My data looks like this:

Col1 Var1 Var2
A     1    NA
B     NA    1
C     1    NA
D     1    1

I want to create a grid plot of missing data just as one can do using the Amelia package (https://www.r-bloggers.com/ggplot-your-missing-data-2/)

However, I find the result pretty ugly. Basically, I want variables in col 1 to be in the x axis and Var1 and Va2 on the Y axis. Like grey when present and black when absent. Does that make sense?

Any advice? I placed the Amelia plot below enter image description here

Upvotes: 2

Views: 273

Answers (2)

G5W
G5W

Reputation: 37641

A base R solution

Dat = t(matrix(as.numeric(is.na(df[,2:3])), nrow=nrow(df)))
rownames(Dat) = names(df)[2:3]
colnames(Dat) = df$Col1

heatmap(Dat, NA, NA, scale="none", col=c("gray", "black"))

Heatmap

Data

df = read.table(text="Col1 Var1 Var2
A     1    NA
B     NA    1
C     1    NA
D     1    1", 
header=TRUE)

Upvotes: 2

markus
markus

Reputation: 26343

Here is a ggplot2 option.

Reshape your data from wide to long first and replace ǸA with 0s (or any other value).

df1_long <- tidyr::gather(replace(df1, is.na(df1), 0), key, value, -Col1)

Now plot

library(ggplot)
ggplot(df1_long, aes(Col1, key, fill = factor(value))) +
  geom_tile() + 
  scale_x_discrete(expand = c(0, 0)) +
  scale_y_discrete(expand = c(0, 0)) +
  scale_fill_manual(values = c(`0` = "black",
                               `1` = "grey80"),
                    labels = c("Missing", "Observed")) +
  labs(title = "Your Title",
       fill = NULL,
       x = NULL,
       y = NULL) +
  coord_equal() +
  theme(legend.position = "bottom")

enter image description here

data

df1 <- structure(list(Col1 = c("A", "B", "C", "D"), Var1 = c(1L, NA, 
1L, 1L), Var2 = c(NA, 1L, NA, 1L)), .Names = c("Col1", "Var1", 
"Var2"), class = "data.frame", row.names = c(NA, -4L))

Upvotes: 1

Related Questions