使用 plot.ts() 的时间序列的日期转换问题

Question

I have a dataframe with date in yyyy-mm-dd format and share prices but I can't see the date in the x axis when plotting it using plot.ts() . I have tried a few alternatives mentioned below but they did not work.

I converted the date using

1. data$Date<-as.Date(data$Date, "%Y-%m-%d") OR
2. data$Date<-ymd(data$Date) using the lubridate package

and then did

bby <- ts(data=data$Share_price, frequency=2, start=c(data[1,"Date"]))
plot.ts(bby))

It was unsuccessful.

I also tried

bby <- ts(data=data$Share_price, frequency=2,
     start=as.Date("2017-10-05"), end=as.Date("2019-10-04"))`

and then plot.ts(bby)

but again it did not work. I always end up with the below graph:

Thanks for your help.

Answer 1

The ts setup is a bit clunky and its plots are not always the friendliest. You could set up your data as follows. Note that you need to specify the 'start' for the time series object below in a rather painful manner, and then (because it only accepts regular time intervals) you lose the day gap in the original series:

z <- seq.Date(as.Date('2017-10-05'), by = 1, length.out = 8)
data <- data.frame(Date = z[-(3:4)],
                   Share_price = c(1708.84, 1718.40, 1724.14, 1762.39, 1766.21, 1813.07))
myts <- ts(data$Share_price,
           start = c(2017, as.numeric(format(data$Date[1], "%j"))),
           frequency = 365)
plot(myts)

The x-axis now has decimal offsets from the year 2017.

Perhaps the xts package is better here? It can handle irregular time indices.

library(xts)
myxts <- xts(data$Share_price, data$Date)
plot(myxts)

See the plot.xts() documentation for all sorts of bells and whistles to embellish your graph.

Answer 2

Normally ts class is used for monthly, quarterly or yearly but not daily data.

Assuming that DF has Date and numeric columns as per reproducible form shown in the Note at the end, convert to zoo class. Then we can easily plot using classic graphics, ggplot2 or lattice graphics.

library(zoo)
z <- read.zoo(DF)

plot(z, main = "classic")

library(ggplot2)
autoplot(z) + 

library(lattice)
xyplot(z)

Plotting all three together using cowplot

library(gridGraphics)
library(cowplot)

plot(z)
p1 <- recordPlot()

library(ggplot2)
p2 <- autoplot(z)

library(lattice)
p3 <- xyplot(z)

plot_grid(p1, p2, p3, labels = c("classic", "ggplot2", "lattice"), nrow = 1)

giving:

Note

DF <- structure(list(Date = structure(c(17444L, 17445L, 17448L, 17449L, 
17450L, 17451L), class = "Date"), Share_price = c(1708.84, 1718.4, 
1724.14, 1762.39, 1766.21, 1813.07)), class = "data.frame", row.names = c(NA, 
-6L))