plot-methods: Methods for 'plot' in Package 'processdata'
In ppstat: Point Process Statistics

Description Usage Arguments Details Value Author(s) See Also Examples

Methods for plotting process data contained in objects of class ContinuousProcess, MarkedPointProcess or JumpProcess.

## S4 method for signature 'ContinuousProcess'
plot(x, nPoints = 200, ...)

## S4 method for signature 'MarkedPointProcess,character'
plot(x, y, nPoints = 200, ...)

## S4 method for signature 'JumpProcess'
plot(x, nPoints = 200, ...)

## S4 method for signature 'ProcessPlotData'
plot(x, ...)

## S4 method for signature 'ContinuousProcess'
getPlotData(object, y = '@bottom', nPoints = 200,
allUnitData = FALSE, selectPoints = NULL, dropLevels = 1, ...)

## S4 method for signature 'MarkedPointProcess,character'
getPlotData(object, y = '@mark', nPoints = 200, allUnitData = FALSE,
allMarkValueData = isTRUE(y %in% names(getMarkValue(object))), ...)

## S4 method for signature 'JumpProcess'
getPlotData(object, nPoints = 200, ...)

`x`	process object to be plotted.
`y`	a `character`. Specifies how the points are plotted for a `MarkedPointProcess` object. Possible values are `'@mark'`, `'@top'`, `'@bottom'`, a name of a continuous process variable, a name of a mark value variable, the unit identification variable name or a numeric. See Details for further explanations. If missing, the value is taken as `'@mark'`.
`object`	process object for which to compute a data set for plotting.
`nPoints`	a `numeric`. Specifies the number of grid points for each process to include in the data set that is plotted. Default value is 200.
`allUnitData`	a `logical`. Specifies if non-process data should be included in the data set for plotting. Default value depends upon whether `y` is a mark value or not.
`allMarkValueData`	a `logical`. Specifies if mark value data should be included in the data set for plotting. Default value is `FALSE`. If
`selectPoints`	a `numeric` vector or `NULL`. Grid points to include in the plot in addition to the `nPoints`.
`dropLevels`	a `numeric` vector or `list` of vectors. Specifies one or more levels to be dropped when plotting factor columns. Default value is 1, which means that the first level is dropped from the plot.
`...`	other arguments.

The underlying plotting engine is ggplot from the ggplot2 package. The result of a call to plot is thus a ggplot object, which can be altered and plotted like any other ggplot object.

The default plot is a plot of the (multivariate) process data held in the process object x divided according to the observation units – the individuals. This is most suitable with relatively few units.

The plot methods rely on the data structure to determine sensible aesthetic mappings of the data for visualization. Hence, numeric columns in a ContinuousProcess object are plotted as continuous sample paths of processes, factor columns are plotted as bars at different levels and event columns in a MarkedPointProcess object are plotted as points. We refer to the former as a ‘Continuous’ type data and the latter two as ‘Track’ data. The plot methods attempt to compute a sensible y-axis to plot either one of the types of data or the combination of ‘Continuous’ and ‘Track’ data.

Modifications to the resulting plot can be done by a combination of arguments to the plot methods or subsequent manipulations of the ggplot2 objects. For instance, with many units and p the value of the plot call, the plot

p + facet_null()

can be a useful alternative to the default. For ContinuousProcess objects with many units and few variables, one can also consider

p + facet_grid(variable ~ .)

while for a MarkedPointProcess a possible alternative is

plot(object, y = "id") + facet_grid(variable ~ ., scale = "free_y")

For JumpProcess objects one can in principle proceed as for a ContinuousProcess objects. However, when plotting the process with discontinuities at the jump times the curve for each unit becomes disconnected. A better alternative then, assuming that the unit identification variable is id, is

p + facet_grid(variable ~ .) + aes(color = id)

See demo(examples, package = "ppstat") for examples. Some aspects of the ggplot objects are not that easily modified after the object has been created. See below for how to deal with some examples.

The y argument controls the y-values used for plotting the point process part of the data set when plotting a MarkedPointProcess object. Values '@top' and '@bottom' place the points above or below, respectively, the plot of the ContinuousProcess part of the object. A numeric value of y, e.g. 0, results in a plot with all points at this particular y-value. A name of a mark value variable places the points at their corresponding mark value and a name of a continuous process variable places the points on the continous process. The default, '@mark', places the points at different levels on the y-axis corresponding to the different marks, and the unit identification variable name places the points at different levels on the y-axis corresponding to the different units.

As descriped above, the factor columns of a process data object are plotted with bars at the different levels of the factor. The dropLevels argument works like the specification of contrasts in a linear model parametrization. That is, the default value of 1 means that the first level is dropped in the plot and bars are plotted for the other levels.

Technically, a call to plot for either of the S4 process data classes above is a two step procedure. First a call to getPlotData that returns the ProcessPlotData object, and then a call of the plot method for that object, which constructs the resulting ggplot2 object. The ProcessPlotData class is, however, currently not exported, nor is the getPlotData methods. The user only needs to understand that additional arguments to plot should be named and are passed on to the getPlotData data extractors and the plot method for the ProcessPlotData object.

The ‘Track’ types of data are plotted as layers on the plot with their own data sets. In general, any argument for geom_point, e.g. shape, used to create the event points layer can be given as an argument of the form point_shape appending ‘point_’ to the argument. Any argument for geom_line, e.g. size, used to create the factor bars layer can be given as an argument of the form factor_size appending ‘factor_’ to the argument. See the examples below.

plot returns a ggplot object.

getPlotData returns an object of class ProcessPlotData, which contains the data that is actually plotted.

Niels Richard Hansen, Niels.R.Hansen@math.ku.dk

ProcessData, ContinuousProcess, MarkedPointProcess, JumpProcess, ProcessPlotData, ggplot

## Examples of plot usages are collected in a demo.
## Not run: 
demo(examples, package = "processdata")
## End(Not run)
## Certain parameters used in the layers for the tracks can be specified
## when calling the plot method. 
## Not run: 
data(example, package = "processdata")
plot(pointExam, point_size = 4, point_shape = 15)
plot(factExam, factor_size = 1, factor_alpha = 0.5)
## End(Not run)