How to plot 2 time series lines and their trend lines in r

Question

I have the code given below. I have data in 3 columns in csv file having columns names as Year, GAC and A1C. I use the code, it just plots time series of only GAC and it also plots its trend line as well as shown in figure

How I can plot both the time series lines of GAC and A1C and their trend lines on the same graph in the code. Please help

h=ggplot(Data, aes(Year, GAC, group=1, xaxt='n', las=2))+
  geom_line()+
  geom_smooth(method=lm,se=FALSE,size= 1.5)

In order for us to help you, please provide a [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). For example, to produce a minimal data set, you can use `head()`, `subset()`, or the indices. Then use `dput()` to give us something that can be put in R immediately. Also, please make sure you know what to do [when someone answers your question](https://stackoverflow.com/help/someone-answers). More info can be found at Stack Overflow's [help center](https://stackoverflow.com/help). — iamericfletcher, Aug 11 '20 at 18:22
You're blending base and `ggplot2` options: `xaxt='n', las=2` don't belong in ggplot. Also, `group=1` should likely be outside `aes(...)`, as it is static and does not depend on anything in the actual data. — r2evans, Aug 11 '20 at 18:24

r2evans · Answer 1 · 2020-08-11T18:31:41.280

ggplot2 really wants data in a "long" format, so I suggest you reshape your data from "wide" (3 columns) to "long" (2 columns). There are multiple ways to do this, SO has many questions about "reshape wide to long", but here is one.

Lacking your data, here is something random.

set.seed(42)
dat <- data.frame(Year = 1900 + 1:50, GAC = cumsum(runif(50)), Unk = cumsum(rnorm(50)))
head(dat)
#   Year      GAC        Unk
# 1 1901 0.914806 -0.4304691
# 2 1902 1.851881 -0.6877385
# 3 1903 2.138021 -2.4509016
# 4 1904 2.968469 -1.9908042
# 5 1905 3.610214 -2.6307991
# 6 1906 4.129310 -2.1753490

# library(tidyr)
head(tidyr::pivot_longer(dat, GAC:Unk))
# # A tibble: 6 x 3
#    Year name   value
#   <dbl> <chr>  <dbl>
# 1  1901 GAC    0.915
# 2  1901 Unk   -0.430
# 3  1902 GAC    1.85 
# 4  1902 Unk   -0.688
# 5  1903 GAC    2.14 
# 6  1903 Unk   -2.45

And a plot:

library(ggplot2)
ggplot(tidyr::pivot_longer(dat, GAC:Unk), aes(Year, value, color = name)) +
  geom_line() +
  geom_smooth(method = lm, se = FALSE, size = 1.5)

There are ways, of course, to do this without reshaping the data. I discourage this for two reasons:

it does not scale well, requiring you to add a new geom_line for each line series you want to plot (as well as geom_smooth if you want it); and
if you want a legend, then it tends to be much easier to allow ggplot to handle that for you. This can only be done (well/easily) when ggplot handles the groups/lines.

ggplot(mapping = aes(x = Year)) +
  geom_line(aes(y = GAC), color = "red", data = dat) +
  geom_smooth(aes(Year, GAC), data = dat, color = "red", method = lm, se = FALSE, size = 1.5) +
  geom_line(aes(y = Unk), color = "blue", data = dat) +
  geom_smooth(aes(Year, Unk), data = dat, color = "blue", method = lm, se = FALSE, size = 1.5)

How to plot 2 time series lines and their trend lines in r

1 Answers1