Display a Third Variable in a X-Y Plot as Percentiles
Displays a third variable on a X-Y plot where the the third variable is represented by symbols indicating within which group defined by the data's 2nd, 5th, 25th, 50th, 75th, 95th and 98th percentiles plotted a data value falls. The colours of the symbols may be optionally changed. The x-y plot axes may be optionally displayed with logarithmic (base 10) scaling. Optionally a legend (two options) may be added to the plot.
Where the x-y data are parts from a composition they should be plotted as symmetric coordinates, and the third variable as a log-ratio if it is from the same composition.
xyplot.eda8(xx, yy, zz, sfact = 1, xlim = NULL, ylim = NULL, xlab = deparse(substitute(xx)), ylab = deparse(substitute(yy)), zlab = deparse(substitute(zz)), main = "", log = NULL, ifgrey = FALSE, symcolr = NULL, iflgnd = FALSE, pctile = FALSE, title = deparse(substitute(zz)), cex.lgnd = 0.8, ...)
xx |
name of the x-axis variable. |
yy |
name of the y-axis variable. |
zz |
name of the third variable to be plotted. |
sfact |
controls the absolute size of the plotted symbols, by default |
log |
if it is required to display the data with logarithmic axis scaling, set |
xlim |
user defined limits for the x-axis, see Details below. |
ylim |
user defined limits for the y-axis, see Details below. |
xlab |
by default the character string for |
ylab |
by default the character string for |
zlab |
by default the character string for |
main |
an alternative plot title, see Details below. |
ifgrey |
set |
symcolr |
the default is a colour plot and default colours are provided, deeper blues for lower values, green for the middle 50% of the data, and oranges and reds for higher values. A set of alternate symbol colours can be provided by defining |
iflgnd |
the default is no legend. If a legend is required set |
pctile |
the default legend displays the range of values each symbol represents. Alternately, the percentiles may be displayed rather than their values by setting |
title |
a short title for the legend, e.g., |
cex.lgnd |
controls the scaling of the legend box and text, but not the symbols so that they match those in the plot, the default is |
... |
further arguments to be passed to methods. For example, if it is required to make the plot title smaller, add |
The selected percentiles, 2nd, 5th, 25th, 50th, 75th, 95th and 98th, divide the data into 8 groups. Values below the median are represented by increasingly larger deeper blue circles below the 25th percentile (Q1), and values above the 75th percentile (Q3) by increasingly larger orange and red squares. The mid 50% of the data are represented by green symbols, circles for the median (Q2) to Q1, and squares for the median (Q2) to Q3.
A summary table of the values of the symbol intervals, the number of values plotting as each symbol, and symbol shapes, sizes and colours is displayed on the current device.
If zlab and main are undefined a default a plot title is generated by appending the input variable name text string to "EDA Percentile Based Plot for ". If no plot title is required set zlab = " ", and if some user defined plot title is required it should be defined in main, e.g. main = "Plot Title Text".
If the grey-scale option is chosen the symbols are plotted 100% black for the far outliers, 85% black for the near outliers, 70% black for values within the whiskers, and 60% black for values falling within the middle 50% of the data.
The default colours, symcolr = c(25, 22, 20, 13, 13, 6, 4, 1), are selected from the rainbow(36) pallette, and alternate colour schemes need to be selected from the same palette. See display.rainbow for the available colours. It is essential that 8 colours be provided, e.g., symcolr = c(27, 24, 22, 12, 12, 5, 3, 36), if exactly 8 are not provided the default colours will be displayed.
For compositional data the function gx.symm.coords.mat is required, and arguments xx and yy are replaced by a call to that function which returms the symmetric coordinated for those parts of the composition, and the axes are labelled appropriately. See example below.
Any less than detection limit values represented by negative values, or zeros or other numeric codes representing blanks in the data, must be removed prior to executing this function, see ltdl.fix.df.
Any data vectors including NAs are removed prior to displaying the plot.
Robert G. Garrett
## Make test data available data(sind) attach(sind) ## Plot a default percentile display xyplot.eda8(Fe, Mn, Zn) ## Plot with more appropriate axis scaling and labelling ## with a user specified title xyplot.eda8(Fe, Mn, Zn, sfact = 2.0, log = "y", xlab = "Fe (pct) in stream sediment", ylab = "Mn (mg/kg) in stream sediment", main = "Howarth & Sinding-Larsen Test Data\nZn (mg/kg)") ## Display a grey-scale equivalent of the above plot xyplot.eda8(Fe, Mn, Zn, sfact = 2, log = "y", ifgrey = TRUE, xlab = "Fe (pct) in stream sediment", ylab = "Mn (mg/kg) in stream sediment", main = "Howarth & Sinding-Larsen Test Data\nZn (mg/kg)") ## Plot the same display with an alternate colour scheme xyplot.eda8(Fe, Mn, Zn, sfact = 2, log = "y", xlab = "Fe (pct) in stream sediment", ylab = "Mn (mg/kg) in stream sediment", main = "Howarth & Sinding-Larsen Test Data\nZn (mg/kg)", symcolr = c(27, 24, 22, 12, 12, 5, 3, 36)) ## Detach test data detach(sind) rm(sind) ## More appropriately for compositional data data(nockolds) attach(nockolds) ## Plot a minor element log ratio against the symmetric coordinates for ## Si and Al xyplot.eda8(gx.symm.coords.mat(nockolds, 1, 2), log(P/Ti)) ## Detach test data detach(nockolds) rm(nockolds)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.