Solved – How to plot multiple loess curves in a xyplot panel data

data visualizationloessr

Imagine the following simplified longitudinal data with 4 individuals and 3 variables (y1, y2, y3) measured at 50 occasions :

dat<-data.frame(subject=as.factor(rep(1:4, each=50)),time=rep(1:50,4),y1=rnorm(200,50,20),y2=rnorm(200,70,20),y3=rnorm(200,20,20))

It is easy to plot at the same time the raw time series with this code :

xyplot(y1+y2+y3~time|subject,dat,type='l')

which result is :

enter image description here

I would like to plot exactly the same, but with loess curves instead of lines. And here is my problem!
The following code :

xyplot(y1+y2+y3~time|subject,dat,
panel=function(x,y)
    {panel.loess(x,y,span=0.2)
})

doesn't work because it considers y1+y2+y3 as a single variable, thus plotting only one loess curve by panel.
How is it possible to plot each loess curve in each panel?

PS. I don't want to use type="smooth" because I need to specify the span of the loess.

Best Answer

Use type='smooth' instead of type='l':

xyplot(y1+y2+y3~time|subject,dat,type='smooth')

Related Solutions

Solved – How to add symbols on smooth lines in a R’s lattice xyplot

Here is a way to achieve this:

library(lattice)
library(gridExtra)

my.df <- data.frame(y=c(y1, y2, y3, z1, z2, z3), 
                    g=gl(6, 50), x=rep(x, 6))

my.panel.loess <- function(x, y, span=2/3, ...) {
  loess.fit <- loess.smooth(x, y, span=span)
  panel.lines(loess.fit$x, loess.fit$y, ...)
}

pp <- list()
pp[[1]] <- xyplot(y ~ x, data=my.df, groups=g, type="b", 
                  col=rep(c("blue","red"), each=3), cex=.6, 
                  panel=function(...) 
                    panel.superpose(panel.groups=my.panel.loess, ...))

pp[[2]] <- update(pp[[1]], span=1/5)
pp[[3]] <- update(pp[[1]], col=1:6, type=c("p", "g"))
pp[[4]] <- update(pp[[1]], pch=rep(c(2,6), each=3))

do.call(grid.arrange,  pp)

You will notice that I had to briefly define a custom panel function for the smoother. This is because I need to pass an extra parameter to panel.lines ("p"+"l"="b"); note that it cannot be used directly when calling panel.lines since this formal parameter is already defined in xyplot.

The rest of the code is pure drawing with varying parameters (all digested through the ... argument). Although I arranged your data in a more convenient data.frame, I believe your original formulation will work with slight adaptation around the call to panel.groups.

enter image description here

However, you can annotate the curves directly with the directlabels package (a lot of examples are available on R-forge). Something along those lines should work fine:

library(directlabels)
pp0 <- xyplot(y ~ x, data=my.df, groups=g, type="smooth") 
direct.label(pp0, "follow.points")

I should note that there is a nice function, labcurve(), in the Hmisc package that does pretty much the same job.

Solved – Plotting and presenting longitudinal data, options

Here's one idea for visualizing your data. Using a variation of small multiples, you could have two charts: one showing the variables' mean with a focus on a variable of interest (including SE or SD), with the second showing the focused variables individual observations.

The following chart shows three variables across 6 time periods, with 50 observations each. The shaded area indicates the SE of the focus variable, which is highlighted similarly in both charts.

Variable Observations

This could be easily scaled up or down based upon your specific needs. It can also be interactive, with a drop-down selection of a variable, and the focus and observations changing to the new item of interest.

EDIT: Here's another example, using your sample data. This shows the mean and SE (disregarding the NA's, since I don't know how you'll handle the calcs).

Median with SE

These were done in Excel, so I have no idea what the equivalent R code would be, but someone else here probably does. To create this chart, you'll have to nudge Excel into doing what you want. For the sample data you provided, I:

Added four calculated rows: Mean, SE, Mean-SE, and 2*SE
Create a line chart with all of your observations as series across your time periods. Format to your taste (it's probably worth some VBA to format everything at once, instead of individually selecting all 150 series). Format the rest of this chart to your preferences.
Copy the chart, and paste it onto the same worksheet.
Using the copy, delete all the series and add the Mean-SE and 2*SE series.
Convert the chart type from Line Chart to Stacked Area Chart.
Format the bottom series (Mean-SE) to No Fill. This should create the appearance of the 2*SE series floating.
Add the Mean Series, and convert it to a Line Chart Type. This will cause it to appear in front of the 2*SE area series.
Format the Chart Area and Plot Area to No Fill.
Using Page Layout > Align > Snap to Grid align the two charts with the second chart on top.

While this looks pretty convoluted, it only takes 10-15 minutes to complete, which if you're like me is much less than trying to learn R.

Best Answer

Related Solutions

Solved – How to add symbols on smooth lines in a R’s lattice xyplot

Solved – Plotting and presenting longitudinal data, options

Related Question