How to layer in ggplot2 with two graphs of different quantities?

Question

I’m starting with programming and I’m trying to plot a graph overlaying data extracted from two dataframes.

The code is as follows::

#Real Gross Domestic Product - GDP
USgdp <- read_csv('GDP.csv')
colnames(USgdp) <- c("Date","GDP")

#ISM PMI Composite Index
USpmi <- Quandl('ISM/MAN_PMI')

#GDP
gGDP <- ggplot(data = USgdp,
        aes(x = Date, y = GDP,
           colour = GDP)) +
        labs(title = "U.S. Real Gross Domestic Product", x = "Date", y = "GDP") +
        theme_minimal() +
        guides(fill = "none", colour = "none") +
  geom_bar(stat = 'Identity')

#ISM PMI
gPMI <- ggplot(data = USpmi,
        aes(x = Date, y = PMI,
           colour = "darkred")) +
        guides(fill = "none", colour = "none") +
    geom_line(size = 1)

The zero for the line graph reading corresponds to the value 50 of the y-axis.

How can I overlay these two charts by aligning their respective zeros as in this example?

I appreciate the help!

Answer 1

Modify the values of the variable to be overwritten (PMI - 50, in this case). To ensure a better interpretation of the data, add a corresponding secondary axis and indicate to sec_axis the contrary transformation:

library(ggplot2)

# Dados de exemplo
set.seed(7269)
dadosGDP <- data.frame(ano = 1950:2020, GDP = sample(-20:20, 71, TRUE))
dadosPMI <- data.frame(ano = 1950:2020, PMI = sample(30:80, 71, TRUE))

ggplot(dadosGDP, aes(ano, GDP)) +
  geom_col(aes(fill = GDP)) +
  geom_line(aes(y = PMI-50, color = "PMI"), dadosPMI) +  # aes(..., color = "PMI") e scale_color_manual para exibir melhor na legenda
  scale_color_manual(NULL, values = "red") +
  scale_y_continuous(sec.axis = sec_axis(~.+50, "PMI"))

But be aware that overlapping two charts with different scales of magnitude is a bad practice of data visualization, which is why ggplot2 does not have an automatic way to do this. Consider composing a chart with separate Plots: Plotted graphs separately in single window