Map cell colors to data values in kableExtra to create a heatmap table

Question

I have a table below and would like to apply ROW-level heat maps.

(1) Any idea how? Right now the heat map is the same for both values in the single row below.

(2) Is there a way to make the header for the group column NOT be angled 90 degrees? Right now all headers are angled but for the group column it be better without angle=90.

here is the rmd file.

---
title: "Untitled"
output: pdf_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)


library(dplyr)
library(kableExtra)
#d = data.frame(group= c("A","b"),cat1=c(2,50),cat2=c(100,2))
d = data.frame(group= c("A"),cat1=c(2),cat2=c(NA))

d = d %>%
    mutate_if(is.numeric, function(x) {
  cell_spec(x, "latex", bold = F, background = spec_color(x,option="C", begin=.5, end = 0.99))
}) 

```



```{r table , echo= FALSE, comment = FALSE, message= FALSE, warning = FALSE, fig.height=3, fig.width = 8} 

kable(
      d, format ="latex",
      caption = "",
      booktabs = T, 
      longtable = T,
      escape = F ,
      align = "c"
      ) %>% kable_styling(latex_options = c(
        "striped", 
        "repeat_header"
        )
       )%>% row_spec( 0,angle = 90)


```

Note: Ideally it'd be good to have this done with the kableExtra functionality so they color schemes match exactley to other kableExtra tables.

eipi10 · Answer 1 · 2018-09-07T00:05:15.847

I'm actually not sure how to get the desired color mapping from spec_color. As an alternative, in the code below I've generated the color mapping using the colorRamp function and a palette-generating function, which can be adjusted as desired. The code also deals with missing values by coloring them light gray.

---
output: pdf_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(viridis)
library(dplyr)
library(kableExtra)

set.seed(2)
d = data_frame(group = sample(LETTERS[1:5], 10, replace=TRUE), cat1=runif(10, 0, 100), cat2=runif(10, 0, 100))
d[5,2] = NA

max.val = max(d[ , sapply(d, is.numeric)], na.rm=TRUE)

#pal.fnc = colorRamp(viridis_pal(option="C")(2))
pal.fnc = colorRamp(c("red", "yellow", "green"))

d = d %>%
    mutate_if(is.numeric, function(x) {
  cell_spec(round(x,1), "latex", bold = F, color=grey(.3),
            background = rgb(pal.fnc(x/max.val) %>% replace(., is.na(.), 200), maxColorValue=255))
}) 

```

```{r table , echo= FALSE, comment = FALSE, message= FALSE, warning = FALSE, fig.height=3, fig.width = 8} 
kable(
      d, format ="latex",
      linesep="",
      caption = "",
      booktabs = T, 
      longtable = T,
      escape = F ,
      align = "c"
      ) %>% kable_styling(latex_options = c(
        "striped", 
        "repeat_header"
        )
       )
```

Any idea how to get a red-yellow-green color pallet with that or preferrable use option="C" palletet from kableExtra? — user3022875, Sep 05 '18 at 21:30
@epipi10 Please see edit. I don't think this works with NAs. — user3022875, Sep 06 '18 at 17:17
See updated code. The `replace` function is used to replace missing color values (due to missing data) with light gray (the value 200 is used for each RGB channel, but you can adjust this to any value between 0 and 255). — eipi10, Sep 07 '18 at 00:05
Considered posting my answer here. An alternative way to handling the NAs is to use `case_when` within dplyr to ignore missing values. `kable` will then be able to leave a white space instead: https://stackoverflow.com/a/53681894/7347699 — Michael Harper, Dec 08 '18 at 11:18

Map cell colors to data values in kableExtra to create a heatmap table

1 Answers1

Linked