index: Visualize Static Data Relationships

In metrumrg: Pharmacometric Tools for Data Preparation, Modeling, Simulation, and Reporting

Description

Cascading diagnostic plotting for data frames and data digests.

Usage

## S3 method for class 'digest'
as.conditioned(x, ...)
## S3 method for class 'isolated'
as.conditioned(x, ...)
## S3 method for class 'keyed'
as.conditioned(x, roles = as.roles(x), ...)
## S3 method for class 'digest'
as.isolated(x, ...)
## S3 method for class 'isolated'
as.isolated(x, ...)
## S3 method for class 'keyed'
as.isolated(x, ...)
## S3 method for class 'digest'
as.motif(x, ...)
## S3 method for class 'keyed'
as.motif(x, ...)
## S3 method for class 'motif'
as.motif(x, ...)
## S3 method for class 'nm'
as.motif(x, ...)
## S3 method for class 'digest'
as.roles(x, ...)
## S3 method for class 'keyed'
as.roles(x, motif = as.motif(x), ...)
## S3 method for class 'conditioned'
index(x, roles = as.roles(x), ...)
## S3 method for class 'digest'
index(x, motif = as.motif(x), roles = as.roles(x, motif = motif), ...)
## S3 method for class 'isolated'
index(x, roles = as.roles(x), ...)
## S3 method for class 'keyed'
index(x, roles = as.roles(x), ...)
## S3 method for class 'nm'
index(x, density = 20, ...)
## S3 method for class 'conditioned'
plot(x, roles = as.roles(x), ...)
## S3 method for class 'digest'
plot(x, motif = as.motif(x), roles = as.roles(x, motif = motif), ...)
## S3 method for class 'isolated'
plot(x, ...)
## S3 method for class 'keyed'
plot(x, roles = as.roles(x), ...)
## S3 method for class 'conditioned'
splom(
	x, 
	data=NULL, 
	roles = as.roles(x), 
	main = "", 
	xlab = "", 
	pscales = 0, 
	...
)
## S3 method for class 'digest'
splom(
	x, 
	data=NULL, 
	motif = as.motif(x), 
	roles = as.roles(x, motif = motif), 
	main = "", xlab = "", 
	pscales = 0, 
	...
)
## S3 method for class 'keyed'
splom(x, data=NULL, roles = as.roles(x), ...)
legacy(x, ...)
legacy(x) <- value
## S3 method for class 'legacy'
format(x,...)

Arguments

x	object of dispatch
roles	named character vector with elements x, y, and z indicating columns to use for axes or for grouping
motif	a list with elements x, y, and z (optional), each of which are character vectors of column names in x, prioritizing alternative usage for plotting
density	integer: how many levels of x to include per panel
main	not passed to splom
data	not passed to splom
xlab	not passed to splom
pscales	not passed to splom
value	assigned to the legacy attribute of x
...	extra arguments passed to other methods

Details

This is an experimental system for generating diagnostic plots. It was designed to work with the result of digest, but can also work with data frames and keyed data frames, etc.

as.conditioned resolves a digest into a longer list of data frames (class conditioned), each of which contain only single levels of key columns not mentioned in roles. as.isolated operates on conditioned or unconditioned keyed data frames, returning a list (class isolated) of data frames each having exactly one non-key column. Default methods are defined for making index plots and scatterplot matrices of objects of class digest, conditioned, isolated, and keyed. The plot methods currently plot both.

For purposes of plotting and conditioning, it is necessary to specify the roles of the columns. A roles object is a named vector with elements x, y, and z. Typically, y is a dependent variable, x is an independent variable, and z is a grouping variable (optional). y may be specified as '.', in which case it represents the non-key variable in context.

When plotting an entire digest all at once, it would be tedious to specify, in advance, roles for every resulting view of the data. A motif is a general strategy for assigning roles. Unlike roles, it is a list. Like roles, it has elements x, y, and (optional) z. Each of these is a character vector of column names, typically consisting only of key columns and '.'. Order is important: supporting functions assign x, y, and z in the order given, removing any assigned columns from consideration in later assignments. For example, time may be appropriate as x where available, and a subject identifier could be the alternative: motif=list(y='.',x=c('time','id')). Methods exist for creating default motifs for a variety of objects. These are converted to roles in context.

The following are generic: as.motif, as.conditioned, index, as.roles, as.isolated.

motif and roles

are aliases for their as. counterparts.

Value

as.conditioned returns a list of objects each with class conditioned, keyed, data.frame. as.isolated returns a list of objects each with class isolated, conditioned, keyed, data.frame. as.motif and motif return an object of class motif with optional elements x, y, and z. as.roles and roles return a named character vector, mapping column usage to roles x, y, and z. plot, splom, and index return lists of trellis objects, possibly with deep nesting. plot.digest gives a list with each element having members splom and index. legacy returns named character; format.legacy returns character. The assignment version of legacy requires named character and causes it to be appended to the legacy attribute of the object.

Author(s)

Tim Bergsma

References

http://metrumrg.googlecode.com

See Also

• as.keyed

• static

• digest

Examples

digest(Theoph,c('Subject','Time'))
head(digest(Theoph,c('Subject','Time')))
## Not run: index(as.keyed(Theoph,'Subject'))

metrumrg documentation built on May 2, 2019, 5:55 p.m.