How to visualize the original data by the values-CodePudding

I have a dataset that looks like

df <- data.frame(
  name = LETTERS[seq(from = 1, to = 10)],
  var1 = sample(1:20, 10),
  var2 = sample(5:25, 10),
  var3 = sample(10:30, 10),
  var4 = sample(15:35, 10),  
  var5 = sample(20:40, 10)
)

How can I present the original data filling the units by the value in ggplot? Like what can be done by the Excel 'conditional formatting' function?

I guess geom_segment is involved here while my columns are a set of variables (they have meaningful names in the 'real' dataset) rather than numeric values so does my rownames (actually the first column). So I am totally lost how to set x, xend, y, yend, etc and make it just the way it is.

Thank you for your help.

CodePudding user response：

You could achieve your desired result via geom_tile by first converting your data to long format using e.g. tidyr::pivot_longer like so:

set.seed(123)

library(ggplot2)
library(tidyr)

df_long <- df |> 
  pivot_longer(-name, names_to = "var")

ggplot(df_long, aes(x = var, y = name, fill = value))  
  geom_tile()  
  geom_text(aes(label = value))  
  scale_fill_gradient(low = "#F7FCF5", high = "#005A32")

CodePudding user response：

library(data.table)

ggplot(melt(setDT(df), id.var="name"),aes(variable, name, fill=value))   
  geom_bin2d()   
  geom_text(aes(label=value))   
  scale_fill_gradient(low = "lightgreen", high = "darkgreen")   
  scale_x_discrete(position = "top")   
  scale_y_discrete(limits=rev)   
  labs(x="", y="")   
  theme(legend.position="none")