Beautiful Radar Chart in R using FMSB and GGPlot Packages

12 Dec

Beautiful Radar Chart in R using FMSB and GGPlot Packages

A radar chart, also known as a spider plot is used to visualize the values or scores assigned to an individual over multiple quantitative variables, where each variable corresponds to a specific axis.

This article describes how to create a radar chart in R using two different packages: the fmsb or the ggradar R packages.

Note that, the fmsb radar chart is an R base plot. The ggradar package builds a ggplot spider plot.

You will learn:

how to create a beautiful fmsb radar chart
how to create ggplot radar chart
alternatives to radar charts

Contents:

Demo data
fmsb radar chart
ggplot radar chart using the ggradar R package
Alternatives to radar charts
- Case when all quantitative variables have the same scale
- Case when you have a lot of individuals to plot or if your variables have different scales
Conclusion

Demo data

We’ll use a demo data containing exam scores for 3 students on 9 topics (Biology, Physics, etc). The scores range from 0 to 20. Columns are quantitative variables and rows are individuals.

# Demo data
exam_scores <- data.frame(
    row.names = c("Student.1", "Student.2", "Student.3"),
      Biology = c(7.9, 3.9, 9.4),
      Physics = c(10, 20, 0),
        Maths = c(3.7, 11.5, 2.5),
        Sport = c(8.7, 20, 4),
      English = c(7.9, 7.2, 12.4),
    Geography = c(6.4, 10.5, 6.5),
          Art = c(2.4, 0.2, 9.8),
  Programming = c(0, 0, 20),
        Music = c(20, 20, 20)
)
exam_scores

##           Biology Physics Maths Sport English Geography Art Programming Music
## Student.1     7.9      10   3.7   8.7     7.9       6.4 2.4           0    20
## Student.2     3.9      20  11.5  20.0     7.2      10.5 0.2           0    20
## Student.3     9.4       0   2.5   4.0    12.4       6.5 9.8          20    20

fmsb radar chart

Prerequisites

Install the fmsb R package:

install.packages("fmsb")

Load the package:

library(fmsb)

Data preparation

The data should be organized as follow:

The row 1 must contain the maximum values for each variable
The row 2 must contain the minimum values for each variable
Data for cases or individuals should be given starting from row 3
The number of columns or variables must be more than 2.

# Define the variable ranges: maximum and minimum
max_min <- data.frame(
  Biology = c(20, 0), Physics = c(20, 0), Maths = c(20, 0),
  Sport = c(20, 0), English = c(20, 0), Geography = c(20, 0),
  Art = c(20, 0), Programming = c(20, 0), Music = c(20, 0)
)
rownames(max_min) <- c("Max", "Min")

# Bind the variable ranges to the data
df <- rbind(max_min, exam_scores)
df

##           Biology Physics Maths Sport English Geography  Art Programming Music
## Max          20.0      20  20.0  20.0    20.0      20.0 20.0          20    20
## Min           0.0       0   0.0   0.0     0.0       0.0  0.0           0     0
## Student.1     7.9      10   3.7   8.7     7.9       6.4  2.4           0    20
## Student.2     3.9      20  11.5  20.0     7.2      10.5  0.2           0    20
## Student.3     9.4       0   2.5   4.0    12.4       6.5  9.8          20    20

Basic radar plot

# Plot the data for student 1
library(fmsb)
student1_data <- df[c("Max", "Min", "Student.1"), ]
radarchart(student1_data)

It can be seen that the Student 1 has a high score in Music and Physics compared to the other topics.

Customize the radar charts

Key arguments to customize the different components of the fmsb radar chart:

Variable options
- vlabels: variable labels
- vlcex: controls the font size of variable labels
Polygon options:
- pcol: line color
- pfcol: fill color
- plwd: line width
- plty: line types. Can be a numeric vector 1:6 or a character vector c(“solid”, “dashed”, “dotted”, “dotdash”, “longdash”, “twodash”). To remove the line, use plty = 0 or plty = “blank”.
Grid options:
- cglcol: line color
- cglty: line type
- cglwd: line width
Axis options:
- axislabcol: color of axis label and numbers. Default is “blue”.
- caxislabels: Character vector to be used as labels on the center axis.

Helper function to produce a beautiful radar chart:

create_beautiful_radarchart <- function(data, color = "#00AFBB", 
                                        vlabels = colnames(data), vlcex = 0.7,
                                        caxislabels = NULL, title = NULL, ...){
  radarchart(
    data, axistype = 1,
    # Customize the polygon
    pcol = color, pfcol = scales::alpha(color, 0.5), plwd = 2, plty = 1,
    # Customize the grid
    cglcol = "grey", cglty = 1, cglwd = 0.8,
    # Customize the axis
    axislabcol = "grey", 
    # Variable labels
    vlcex = vlcex, vlabels = vlabels,
    caxislabels = caxislabels, title = title, ...
  )
}

In the code above, we used the function alpha() [in scales package] to change the polygon fill color transparency.

# Reduce plot margin using par()
op <- par(mar = c(1, 2, 2, 1))
create_beautiful_radarchart(student1_data, caxislabels = c(0, 5, 10, 15, 20))
par(op)

Create radar charts for multiple individuals

Create the radar chart of the three students on the same plot:

# Reduce plot margin using par()
op <- par(mar = c(1, 2, 2, 2))
# Create the radar charts
create_beautiful_radarchart(
  data = df, caxislabels = c(0, 5, 10, 15, 20),
  color = c("#00AFBB", "#E7B800", "#FC4E07")
)
# Add an horizontal legend
legend(
  x = "bottom", legend = rownames(df[-c(1,2),]), horiz = TRUE,
  bty = "n", pch = 20 , col = c("#00AFBB", "#E7B800", "#FC4E07"),
  text.col = "black", cex = 1, pt.cex = 1.5
  )
par(op)

Create separated spider charts for each individual. This is recommended when you have more than 3 series.

# Define colors and titles
colors <- c("#00AFBB", "#E7B800", "#FC4E07")
titles <- c("Student.1", "Student.2", "Student.3")

# Reduce plot margin using par()
# Split the screen in 3 parts
op <- par(mar = c(1, 1, 1, 1))
par(mfrow = c(1,3))

# Create the radar chart
for(i in 1:3){
  create_beautiful_radarchart(
    data = df[c(1, 2, i+2), ], caxislabels = c(0, 5, 10, 15, 20),
    color = colors[i], title = titles[i]
    )
}
par(op)

Compare every profile to an average profile

Radar charts are most useful if the profile of every individual is compared to an average profile.

Create a demo data containing exam scores for 10 students:

set.seed(123)
df <- as.data.frame(
  matrix(sample(2:20 , 90 , replace = TRUE),
         ncol=9, byrow = TRUE)
  )
colnames(df) <- c(
  "Biology", "Physics", "Maths", "Sport", "English", 
  "Geography", "Art", "Programming", "Music"
  )
rownames(df) <- paste0("Student.", 1:nrow(df))
head(df)

##           Biology Physics Maths Sport English Geography Art Programming Music
## Student.1      16      20    15     4      11        19  12           6    15
## Student.2       6      20    10     4       9         8  11          10    20
## Student.3       5      15    18    12       8        13  16          11    14
## Student.4       8      10    10    11       8         7   3           6     9
## Student.5      13      14    19     2       7        16  10          16    17
## Student.6       7      12     9     8      17        18  19          18     3

Rescale each variable to range between 0 and 1:

library(scales)
df_scaled <- round(apply(df, 2, scales::rescale), 2)
df_scaled <- as.data.frame(df_scaled)
head(df_scaled)

##           Biology Physics Maths Sport English Geography  Art Programming Music
## Student.1    1.00    1.00  0.69  0.11    0.47      1.00 0.56        0.00  0.71
## Student.2    0.09    1.00  0.31  0.11    0.35      0.27 0.50        0.33  1.00
## Student.3    0.00    0.69  0.92  0.56    0.29      0.60 0.81        0.42  0.65
## Student.4    0.27    0.38  0.31  0.50    0.29      0.20 0.00        0.00  0.35
## Student.5    0.73    0.62  1.00  0.00    0.24      0.80 0.44        0.83  0.82
## Student.6    0.18    0.50  0.23  0.33    0.82      0.93 1.00        1.00  0.00

Prepare the data for creating the radar plot using the fmsb package:

# Variables summary
# Get the minimum and the max of every column  
col_max <- apply(df_scaled, 2, max)
col_min <- apply(df_scaled, 2, min)
# Calculate the average profile 
col_mean <- apply(df_scaled, 2, mean)
# Put together the summary of columns
col_summary <- t(data.frame(Max = col_max, Min = col_min, Average = col_mean))


# Bind variables summary to the data
df_scaled2 <- as.data.frame(rbind(col_summary, df_scaled))
head(df_scaled2)

##           Biology Physics Maths Sport English Geography   Art Programming Music
## Max         1.000   1.000 1.000 1.000   1.000     1.000 1.000        1.00 1.000
## Min         0.000   0.000 0.000 0.000   0.000     0.000 0.000        0.00 0.000
## Average     0.464   0.575 0.476 0.427   0.423     0.587 0.544        0.50 0.629
## Student.1   1.000   1.000 0.690 0.110   0.470     1.000 0.560        0.00 0.710
## Student.2   0.090   1.000 0.310 0.110   0.350     0.270 0.500        0.33 1.000
## Student.3   0.000   0.690 0.920 0.560   0.290     0.600 0.810        0.42 0.650

Produce radar plots showing both the average profile and the individual profile:

opar <- par() 
# Define settings for plotting in a 3x4 grid, with appropriate margins:
par(mar = rep(0.8,4))
par(mfrow = c(3,4))
# Produce a radar-chart for each student
for (i in 4:nrow(df_scaled2)) {
  radarchart(
    df_scaled2[c(1:3, i), ],
    pfcol = c("#99999980",NA),
    pcol= c(NA,2), plty = 1, plwd = 2,
    title = row.names(df_scaled2)[i]
  )
}
# Restore the standard par() settings
par <- par(opar)

ggplot radar chart using the ggradar R package

Prerequisites

Installation:

devtools::install_github("ricardo-bion/ggradar")

Loading the package:

library("ggradar")

Key function and arguments

ggradar(
  plot.data, values.radar = c("0%", "50%", "100%"),
  grid.min = 0, grid.mid = 0.5, grid.max = 1, 
  )

plot.data: data containing one row per individual or group
values.radar: values to show at minimum, average and maximum grid lines
grid.min: value at which minimum grid line is plotted
grid.mid: value at which average grid line is plotted
grid.max: value at which maximum grid line is plotted

Data preparation

All variables in the data should be at the same scale. If this is not the case, you need to rescale the data.

For example, you can rescale the variables to have a minimum of 0 and a maximum of 1 using the function rescale() [scales package]. We’ll describe this method in the next sections.

library(tidyverse)
# Put row names into  a column named group
df <- exam_scores %>% rownames_to_column("group")
df

##       group Biology Physics Maths Sport English Geography Art Programming Music
## 1 Student.1     7.9      10   3.7   8.7     7.9       6.4 2.4           0    20
## 2 Student.2     3.9      20  11.5  20.0     7.2      10.5 0.2           0    20
## 3 Student.3     9.4       0   2.5   4.0    12.4       6.5 9.8          20    20

Basic radar plot

# Plotting student 1
ggradar(
  df[1, ], 
  values.radar = c("0", "10", "20"),
  grid.min = 0, grid.mid = 10, grid.max = 20
  )

Customize radar charts

Key arguments to customize the different components of the ggplot radar chart. For more options see the documentation.

ggradar(
  df[1, ], 
  values.radar = c("0", "10", "20"),
  grid.min = 0, grid.mid = 10, grid.max = 20,
  # Polygons
  group.line.width = 1, 
  group.point.size = 3,
  group.colours = "#00AFBB",
  # Background and grid lines
  background.circle.colour = "white",
  gridline.mid.colour = "grey"
  )

Radar chart with multiple individuals or groups

Create the radar chart of the three students on the same plot:

ggradar(
  df, 
  values.radar = c("0", "10", "20"),
  grid.min = 0, grid.mid = 10, grid.max = 20,
  # Polygons
  group.line.width = 1, 
  group.point.size = 3,
  group.colours = c("#00AFBB", "#E7B800", "#FC4E07"),
  # Background and grid lines
  background.circle.colour = "white",
  gridline.mid.colour = "grey",
  legend.position = "bottom"
  )

Alternatives to radar charts

A circular plot is difficult to read. An alternative to a radar chart is an ordered lolliplot or dotchart. This section describes how to create dotcharts. The ggpubr R package will be used in this section to create a dotchart.

Load required packages:

library(tidyverse)
library(ggpubr)

Case when all quantitative variables have the same scale

Displaying one individual

Data preparation:

df2 <- t(exam_scores) %>%
  as.data.frame() %>%
  rownames_to_column("Field")
df2

##         Field Student.1 Student.2 Student.3
## 1     Biology       7.9       3.9       9.4
## 2     Physics      10.0      20.0       0.0
## 3       Maths       3.7      11.5       2.5
## 4       Sport       8.7      20.0       4.0
## 5     English       7.9       7.2      12.4
## 6   Geography       6.4      10.5       6.5
## 7         Art       2.4       0.2       9.8
## 8 Programming       0.0       0.0      20.0
## 9       Music      20.0      20.0      20.0

Plot creation:

ggdotchart(
  df2, x = "Field", y = "Student.1",
  add = "segments", sorting = "descending",
  ylab = "Exam Score", title = "Student 1"
  )

Displaying two individuals

Data preparation:

df3 <- df2 %>%
  select(Field, Student.1, Student.2) %>%
  pivot_longer(
    cols = c(Student.1, Student.2),
    names_to = "student",
    values_to = "value"
  )
head(df3)

## # A tibble: 6 x 3
##   Field   student   value
##   <chr>   <chr>     <dbl>
## 1 Biology Student.1   7.9
## 2 Biology Student.2   3.9
## 3 Physics Student.1  10  
## 4 Physics Student.2  20  
## 5 Maths   Student.1   3.7
## 6 Maths   Student.2  11.5

Plot creation:

ggdotchart(
  df3, x = "Field", y = "value", 
  group = "student", color = "student", palette = "jco",
  add = "segment", position = position_dodge(0.3),
  sorting = "descending"
  )

Displaying multiple individuals

Data preparation:

df4 <- df2 %>%
  select(Field, Student.1, Student.2, Student.3) %>%
  pivot_longer(
    cols = c(Student.1, Student.2, Student.3),
    names_to = "student",
    values_to = "value"
  )
head(df4)

## # A tibble: 6 x 3
##   Field   student   value
##   <chr>   <chr>     <dbl>
## 1 Biology Student.1   7.9
## 2 Biology Student.2   3.9
## 3 Biology Student.3   9.4
## 4 Physics Student.1  10  
## 5 Physics Student.2  20  
## 6 Physics Student.3   0

Plot creation:

ggdotchart(
  df4, x = "Field", y = "value", 
  group = "student", color = "student", palette = "jco",
  add = "segment", position = position_dodge(0.3),
  sorting = "descending", facet.by = "student",
  rotate = TRUE, legend = "none"
  )

Case when you have a lot of individuals to plot or if your variables have different scales

A solution is to create a parallel coordinates plot.

library(GGally)
ggparcoord(
  iris,
  columns = 1:4, groupColumn = 5, order = "anyClass",
  showPoints = TRUE, 
  title = "Parallel Coordinate Plot for the Iris Data",
  alphaLines = 0.3
  ) + 
  theme_bw() +
  theme(legend.position = "top")

Note that, the default of the function ggparcoord() is to rescale each variable by subtracting the mean and dividing by the standard deviation.

Conclusion

This article describes how to create radar chart in R for one or multiple individuals using the fmsb package and the ggradar package (a ggplot2 extension).