Welcome to OStack Knowledge Sharing Community for programmer and developer-Open, Learning and Share
Welcome To Ask or Share your Answers For Others

Categories

0 votes
639 views
in Technique[技术] by (71.8m points)

r - Date format for plotting x axis ticks of time series data

The data files have date is the format i.e. 1975M1, 1975M2, ... 2011M12 for time series data. when plotting this data using R, I want the x-axis to display the months on tick axis.

For the dates to be read properly, I have tried replacing the M by - to get %Y-%m format but that doesnt seem good for drawTimeAxis from hydroTSM package which perhaps requires %Y-%M-%d format. It gives error that incorrect number of dimensions for ticks dimension.

Another method of parsing and formatting the data as in x$newdate <- strptime(as.character(x$date), "%Y-%m") and then format(x$newdate,""%Y-%m") also doesnt read the date and gives error... all NA.

date <- as.Date(data[,1] the error that character string is not in a standard unambiguous format and ts <- read.zoo(xts, as.yearmon(x[,1])) gives bad enries at data rows.

Please give solution of how this data can be read with the date information.

A small subset of the data file

date    x   x2
1975M1  112.44  113.12
1975M2  113.1   114.36
1975M3  115.04  114.81
1975M4  117.65  115.35
1975M5  119.5   116.92
1975M6  121.4   118.56
1975M7  120.64  118.97
1975M8  119.12  119.84
1975M9  118.91  120.59
1975M10 120.58  122.3
1975M11 121.26  123.35
1975M12 122.34  123.33

Update: The answers so far solve the problem of reading the date correctly by using %YM%m in the xts package or adding the day for getting standard format. The customizing of tick axis is still a problem. The drawTimeAxis is giving dimension error and plot commands are not showing monthly labels for more than one year of data or otherwise . Any methods of customizing tick axis ?

See Question&Answers more detail:os

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome To Ask or Share your Answers For Others

1 Answer

0 votes
by (71.8m points)

Perhaps you are not using as.yearmon() correctly, because the following works for me (using dat from Gavin's answer):

library(zoo)
dat$date <- as.yearmon(dat$date, "%YM%m")

Thus, working through to getting things to plot correctly:

  1. Your data:

    dat <- read.table(text = "date    x   x2
    1975M1  112.44  113.12
    1975M2  113.1   114.36
    1975M3  115.04  114.81
    1975M4  117.65  115.35
    1975M5  119.5   116.92
    1975M6  121.4   118.56
    1975M7  120.64  118.97
    1975M8  119.12  119.84
    1975M9  118.91  120.59
    1975M10 120.58  122.3
    1975M11 121.26  123.35
    1975M12 122.34  123.33", header = TRUE)
    
  2. Conversion to xts using as.yearmon() from the "zoo" package.

    library(xts) # Will also load zoo
    dat.xts <- xts(dat[-1], 
                   order.by = as.yearmon(dat$date, "%YM%m"))
    dat.xts
    #               x     x2
    # Jan 1975 112.44 113.12
    # Feb 1975 113.10 114.36
    # Mar 1975 115.04 114.81
    # Apr 1975 117.65 115.35
    # May 1975 119.50 116.92
    # Jun 1975 121.40 118.56
    # Jul 1975 120.64 118.97
    # Aug 1975 119.12 119.84
    # Sep 1975 118.91 120.59
    # Oct 1975 120.58 122.30
    # Nov 1975 121.26 123.35
    # Dec 1975 122.34 123.33
    
  3. Plotting your data:

    plot.zoo(dat.xts)
    

    enter image description here

    plot.zoo(dat.xts, 
             plot.type="single", 
             col = c("red", "blue"))
    

    enter image description here

Update: Specifying your own axes

Here is some sample data to work with (it's usually nice to share such sample data when asking questions on SO since it makes it easier for others to replicate and address your problem(s)). Note that for this example, we've skipped using the "xts" package since it's not really necessary.

set.seed(1)
dat <- data.frame(date = paste0(rep(1975:1977, each = 12), 
                                "M", rep(1:12, times = 3)),
                  x1 = runif(36, min = 100, max = 140),
                  x2 = runif(36, min = 100, max = 140))
library(zoo) # xts is actually unnecessary if this is all you're doing
# Convert your data to a `zoo` object
dat.z <- zoo(dat[-1], order.by = as.yearmon(dat$date, "%YM%m"))

This is the default plot obtained with plot(dat.z, screen = 1, col = 1:2):

enter image description here

From your comments, it sounds like you want something like monthly labels.

  1. Plot the data, but suppress the x-axis with xaxt = "n"

    plot(dat.z, screen = 1, col = 1:2, xaxt = "n")
    
  2. Do some setup work to have a label for every month. (See ?plot.zoo, from where this is modified.)

    tt <- time(dat.z)
    # The following is just the sequence 1:36. 
    #   If you wanted only every third month plotted,
    #   use a sequence like ix <- seq(1, length(tt), 3)
    ix <- seq_along(tt) 
    # What format do you want for your labels.
    #   This yields abbreviated month - abbreviated year
    fmt <- "%b-%y" 
    labs <- format(tt, fmt) # Generate the vector of your labels
    
  3. Add your axis to your plot. Some experimentation might be needed to find the right sizes for everything. las = 2 makes the labels perpendicular to the axis, which is required if you really feel the need to include a label for every month of each year.

    axis(side = 1, at = tt[ix], labels = labs[ix], 
         tcl = -0.7, cex.axis = 0.7, las = 2)
    

Here is the final plot:

enter image description here

By the way, if you are getting dates like 1977.15 and so on, you might want to read through some of the answers to this question, for example, looking at @joran's use of pretty().


与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome to OStack Knowledge Sharing Community for programmer and developer-Open, Learning and Share
Click Here to Ask a Question

...