ggplot2 (Barplot + LinePlot) - Dual Y axis

Question

I am having a really hard time recreating an excel example with ggplot2. I have tried numerous examples but for some reason I cannot reach my desired result. Can someone please have a look at my example?

df <- structure(list(OccuranceCT = c(4825, 9063, 10635, 8733, 5594, 
2850, 1182, 376, 135, 30, 11), TimesReshop = structure(1:11, .Label = c("1x", 
"2x", "3x", "4x", "5x", "6x", "7x", "8x", "9x", "10x", "11x"), class = "factor"), 
    AverageRepair_HrsPerCar = c(7.48951898445596, 6.50803925852367, 
    5.92154446638458, 5.5703551356922, 5.38877037897748, 5.03508435087719, 
    4.92951776649746, 4.83878377659575, 4.67829259259259, 4.14746333333333, 
    3.54090909090909)), .Names = c("OccuranceCT", "TimesReshop", 
"AverageRepair_HrsPerCar"), row.names = c(NA, 11L), class = "data.frame")

My plot so far:

Plot <- ggplot(df, aes(x=TimesReshop, y=OccuranceCT)) +
  geom_bar(stat = "identity", color="red", fill="#C00000") +
  labs(x = "Car Count", y = "Average Repair Per Hour") + 
  geom_text(aes(label=OccuranceCT), fontface="bold", vjust=1.4, color="black", size=4) +
  theme_minimal()

Plot

This is what I got so far:

And what I am trying to achieve is:

I would be grateful to learn how to add the secondary axis and combine a bar plot with a line plot.

Someone has to give the "dual y-axes are bad" lecture so it may as well be me :) They were not implemented in ggplot2 until recently and the answer by @GGamba shows how to do it. A more ggplot2 way would be to reshape your data from wide-long and plot each variable in its own facet. — neilfws
– neilfws, Commented Jun 19, 2017 at 23:13
Can you please explain how to reshape the data from wide-long? I would like to know how to do it they way you described. Thank you Neil! — Cosmin
– Cosmin, Commented Jun 19, 2017 at 23:25

GGamba · Accepted Answer · 2017-06-19 23:07:31Z

8

ggplot2 supports dual axis (for good or for worse), where the second axis is a linear transformation of the main axis.

We can work it out for this case:

library(ggplot2)
ggplot(df, aes(x = TimesReshop)) +
  geom_col(aes( y = OccuranceCT, fill="redfill")) +
  geom_text(aes(y = OccuranceCT, label = OccuranceCT), fontface = "bold", vjust = 1.4, color = "black", size = 4) +
  geom_line(aes(y = AverageRepair_HrsPerCar * 1500, group = 1, color = 'blackline')) +
  geom_text(aes(y = AverageRepair_HrsPerCar * 1500, label = round(AverageRepair_HrsPerCar, 2)), vjust = 1.4, color = "black", size = 3) +
  scale_y_continuous(sec.axis = sec_axis(trans = ~ . / 1500)) +
  scale_fill_manual('', labels = 'Occurance', values = "#C00000") +
  scale_color_manual('', labels = 'Time Reshop', values = 'black') +
  theme_minimal()

answered Jun 19, 2017 at 23:07

GGamba

13.7k3 gold badges41 silver badges51 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Cosmin Over a year ago

Wow that was fast. Thank you so much GGamba. This is exactly what I was looking for!!! Can you please explain why are you multiplying by 1500 and divide by the same number. How can I figure out by how much do I need to multiply for another example, in the future?

neilfws · Accepted Answer · 2017-06-19 23:33:15Z

5

This answer is in reply to your comment, not to the original question.

Reshaping from wide to long means that we have one column for the dependent variables (OccuranceCT, AverageRepair_HrsPerCar) and another for their values. We could then plot each as bars, in their own facet, like this:

library(tidyr)
library(ggplot2)

df %>% 
  gather(variable, value, -TimesReshop) %>% 
  ggplot(aes(TimesReshop, value)) + 
    geom_col() + 
    facet_grid(variable ~ ., scales = "free")

This allows for quick visual comparison of the variables without the potentially-misleading interpretations that can arise from putting different variables with quite different values in the same plot.

answered Jun 19, 2017 at 23:33

neilfws

34k5 gold badges58 silver badges72 bronze badges

1 Comment

Cosmin Over a year ago

Neil, now I completely understand. Thank you for simplifying and showing me how to do this. I see why you suggested this approach.

Collectives™ on Stack Overflow

ggplot2 (Barplot + LinePlot) - Dual Y axis

2 Answers 2

1 Comment

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related