Plot two lines on one graph in ggplot and R


## 
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

Create dataframe

Let’s create a sample dataframe 

nmonths = 24
x = seq(as.Date("2015/1/1"), by = "month", length.out = nmonths)

prescription1 <- data.frame(
  x,
  Percent.Change = 25 + runif(nmonths,1,100)
)

prescription2 <- data.frame(
  x,
  Percent.Change = 75 + runif(nmonths,1,50)
)

cols = c("dates", "Difference")
colnames(prescription1) = cols
colnames(prescription2) = cols

Straightforward

p = ggplot() + 
  geom_line(data = prescription1, aes(x = dates, y = Difference), color = "blue") +
  geom_line(data = prescription2, aes(x = dates, y = Difference), color = "red") +
  xlab('Dates') +
  ylab('percent.change')

print(p)

Second method using melt()

Using the melt() function, from the reshape package provides a number of advantages.

  • We may plot any number of lines on the plot
  • A legend is provided by default
  • each line has a different colour
  • Only one call to ggplot is necessary

In essence melting converts a dataframe from a “short and wide format into a long and thin format ( more rows )”

print(head(prescription1))
##        dates Difference
## 1 2015-01-01  117.35197
## 2 2015-02-01   82.06761
## 3 2015-03-01   48.21552
## 4 2015-04-01  107.49036
## 5 2015-05-01  105.70461
## 6 2015-06-01   30.87076
prescription = merge(prescription1, prescription2, by="dates")
head(prescription)
##        dates Difference.x Difference.y
## 1 2015-01-01    117.35197    113.98116
## 2 2015-02-01     82.06761    104.97214
## 3 2015-03-01     48.21552     80.61535
## 4 2015-04-01    107.49036     89.48244
## 5 2015-05-01    105.70461     94.89909
## 6 2015-06-01     30.87076    102.23722
prescriptionMelted <- reshape2::melt(prescription, id.var='dates')
head(prescriptionMelted)
##        dates     variable     value
## 1 2015-01-01 Difference.x 117.35197
## 2 2015-02-01 Difference.x  82.06761
## 3 2015-03-01 Difference.x  48.21552
## 4 2015-04-01 Difference.x 107.49036
## 5 2015-05-01 Difference.x 105.70461
## 6 2015-06-01 Difference.x  30.87076
ggplot(prescriptionMelted, aes(x=dates, y=value, col=variable)) + geom_line()

Third Method, using dplyr

We use dplyr to create an extra columns and pipe the results to a combined data frame. Neat and succinct, as so often with dplyr

df1 <- data.frame(dates = x,Variable = rnorm(mean = 0.75,nmonths))
df2 <- data.frame(dates = x,Variable = rnorm(mean = -0.75,nmonths))

df3 <- df1 %>%  mutate(Type = 'Amocycillin') %>%
       bind_rows(df2 %>%
           mutate(Type = 'Penicillin'))


ggplot(df3,aes(y = Variable,x = dates,color = Type)) + 
  geom_line() +
  ggtitle("Merged datasets")


References


http://www.r-cran.com
https://www.rstudio.com
http://www.rpubs.com

https://rviews.rstudio.com
http://www.tendron.net

Alan Brown, CTO, Tendron Systems Ltd

– – – – – – – – – – – –