How to group multiple columns together and plot a bar graph?

Question

I have a data frame that I have split using the splitstackshape package. After splitting I unable to proceed with grouping multiple columns and plotting a bar chart. The code is as follows,

library(tidyverse)
library(splitstackshape)
df <- data.frame(countries=(c("England","Australia,Pakistan", "India,England","Denmark", "",
                             "Australia, Pakistan, New Zealand, England", "United States, England,Pakistan")))
data_split <- splitstackshape::cSplit(df, "countries", ",")
data_split

The output is as follows,

countries_1       countries_2 countries_3     countries_4
1:       England        <NA>        <NA>        <NA>
2:     Australia    Pakistan        <NA>        <NA>
3:         India     England        <NA>        <NA>
4:       Denmark        <NA>        <NA>        <NA>
5:          <NA>        <NA>        <NA>        <NA>
6:     Australia    Pakistan New Zealand     England
7: United States     England    Pakistan        <NA>

With the above output I wish to plot a bar chart containing the frequency of countries in descending order. The sample output is as follows, bar chart showing frequency of countries in descending order

Maybe it is more straightforward to use `separate_rows` from `tidyr` package (https://tidyr.tidyverse.org/reference/separate_rows.html) and then `ggplot` together with `geom_bar` (https://stackoverflow.com/a/37622400/997979) — iago, Jun 10 '20 at 11:31

Ahorn · Accepted Answer · 2020-06-10T10:34:51.353

1

Like this:

library(tidyverse)
library(ggplot2)

df %>% 
  separate_rows(countries, sep = ",") %>% 
  count(countries) %>% 
  ggplot(aes(y = fct_reorder(countries, n), x = n)) +
  geom_col()

Edit based on comment: plot only 10 most common countries:

df %>% 
  separate_rows(countries, sep = ",") %>% 
  count(countries) %>% 
  slice_max(n, n = 10) %>% 
  ggplot(aes(y = fct_reorder(countries, n), x = n)) +
  geom_col()

edited Jun 10 '20 at 10:34

answered Jun 10 '20 at 10:20

Ahorn

3,686
1
10
17

Got it Sir. But the actual data set I am working on has 120 odd countries. So, when I run the output the entire chart is completely overlapped. I wish to show only top 10 countries because the frequency of countries after the first 10 is very less. How can I incorporate that step into the code ? – Silent_bliss Jun 10 '20 at 10:31
Add this line `slice_max(n, n = 10) %>% ` the line with `count` (see updated answer) – Ahorn Jun 10 '20 at 10:35
Thank you so for answering my first question. Means a lot. – Silent_bliss Jun 10 '20 at 10:39
You're welcome! Consider accepting my answer by clicking the checkmark next to it if it answered you question. – Ahorn Jun 10 '20 at 10:42

How to group multiple columns together and plot a bar graph?

1 Answers1