0

I have some time series data and used autoplot function in R to plot my time series. I would like add vertical lines to the plot and text. For example, a line between 2003-2010 with text "train data", 2010-2015 "test" data and 2015- 2018 "predictions". I did it in ordinary R plot, but it is not fancy. I would like to do it with ggplot or autoplot. My code is as follows.

data <- runif(180,100,1000)
ts.data <- ts(data, frequency = 12, start = c(2003,1))
autoplot(ts.data, xlab="Year", ylab="Number of Tourists")

Thanks in advance

Günal
  • 751
  • 1
  • 14
  • 29

2 Answers2

2

You could use the geomtextpath package with geom_textvline to add vertical lines with labels to autoplot since it is a ggplot object. You have to make sure your xintercept is the right number because your data is in dates. Here is some reproducible code:

library(ggplot2)
library(ggfortify)
library(geomtextpath)
autoplot(ts.data, xlab="Year", ylab="Number of Tourists") +
  geom_textvline(xintercept = as.numeric(as.Date(c("2005-01-01"))),
                 label = "train data", vjust = 1.3, col = "blue") +
  geom_textvline(xintercept = as.numeric(as.Date(c("2012-01-01"))),
                 label = "test data", vjust = 1.3, col = "blue") +
  geom_textvline(xintercept = as.numeric(as.Date(c("2017-01-01"))),
                 label = "Predictions", vjust = 1.3, col = "blue")

Created on 2023-02-18 with reprex v2.0.2

Quinten
  • 35,235
  • 5
  • 20
  • 53
2

You can also use ggplot2 itself to get it done:

library(ggplot2)
data <- runif(180,100,1000)
ts.data <- ts(data, frequency = 12, start = c(2003,1))
autoplot(ts.data, xlab="Year", ylab="Number of Tourists") +
  geom_vline(xintercept = as.Date(c("2003-01-01", "2010-01-01", "2015-01-01", "2018-01-01") ), color = "blue")+
  geom_text(aes(x = as.Date("2007-01-01"), y = 1020, label="train data"))+
  geom_text(aes(x = as.Date("2013-01-01"),  y = 1020, label="test"))+
  geom_text(aes(x = as.Date("2016-07-01"),  y = 1020, label="predictions"))

enter image description here

I understand you asked about adding vertical lines, just wonder if it is better to use horizontal lines:

autoplot(ts.data, xlab="Year", ylab="Number of Tourists") +
  geom_segment(aes(x = as.Date("2003-01-01") , y = 1000, xend = as.Date("2010-01-01"), yend = 1000), color = "red") +
  geom_segment(aes(x = as.Date("2010-01-01") , y = 1000, xend = as.Date("2015-01-01"), yend = 1000), color = "blue") +
  geom_segment(aes(x = as.Date("2015-01-01") , y = 1000, xend = as.Date("2018-01-01"), yend = 1000), color = "green") +
  geom_text(aes(x = as.Date("2007-01-01"), y = 1020, label="train data"), color = "red")+
  geom_text(aes(x = as.Date("2013-01-01"),  y = 1020, label="test"), color = "blue")+
  geom_text(aes(x = as.Date("2017-01-01"),  y = 1020, label="predictions"), color="green")

enter image description here

Zhiqiang Wang
  • 6,206
  • 2
  • 13
  • 27
  • Thanks. Can we change the text font? I think it is bold and does not look good. – Günal Feb 18 '23 at 10:58
  • 1
    Yes, definitely. You can change font with `geom_text(..., family =?, fontface =?, size = ?)`. check `geom_text()` document for details. – Zhiqiang Wang Feb 18 '23 at 11:23
  • 2
    the reason why both segment and text are "bold" is that they are being drawn as many times as the data frame has rows. Better to use `annotate(geom = "text(or:segment)")` instead – tjebo Feb 18 '23 at 11:43