R ggplot2 autoplot() function. What's wrong?

Question

I have a dataset "bc" with 2285 observations two variables: "Date" and "Price".

 'data.frame':  2285 obs. of  2 variables:
  $ Date : Date, format: "2017-12-14" "2017-12-13" ...
  $ Price: num  16234 16250 16650 16470 14691 ...

I tried to create a time series object as:

  tsbc <- ts(bc)

Then, I used:

  autoplot(tsbc)

And I get figure below: However, the plot is not how it is supposed to be. Could you help me understand why?

Answer 1

Here is a solution based on xts :

library(ggplot2)

# Generate a dataset
set.seed(1)
bc <- data.frame(Date=seq(as.Date("2016/1/1"), as.Date("2017/12/14"), "days"),
                 Price= cumsum(rnorm(714)))
#   'data.frame':   714 obs. of  2 variables:
# $ Date : Date, format: "2016-01-01" "2016-01-02" ...
# $ Price: num  -0.626 -0.443 -1.278 0.317 0.646 ...

library(xts)
tsbc <- xts(bc$Price, order.by=bc$Date)
autoplot(tsbc)

Otherwise, using ts :

tsbc <- ts(bc$Price, start=c(2016,1), frequency=365)
autoplot(tsbc) + scale_x_yearmon(n=5)

Answer 2

The problem is how you created that time series object tsbc . You are actually creating two time series. And since dates are just numbers under the hood with a class attribute, they lose their attribute when you call ts . That's why autoplot draws a line for Date as well which plots the doubles representing the respective date (see ?base::Dates for more details). Which is obviously not what you want. See @Marco's answer for how to construct the ts object.

However, you don't even need to do that. Why not simply

library(ggplot2)
ggplot(bc, aes(Date, Price)) + geom_line()