lessR

Updates

Chart(), X() and XY(): Parameter type renamed to form, which names the perceptual form of the visualization within the analytic view, such as form="bars" for Chart(). The old name type is still accepted, now with a message that announces the new name form.

Chart(), X() and XY(): The variable arguments x, y, by, and facet now accept expressions, such as Chart(cut(rnorm(50), 3)), X(log(Salary)), or XY(Years, Salary, by=cut(Pre, 2)). An expression evaluates first against the columns of the data frame, then against the user's workspace, with the expression text as the variable label. A vector of variable names such as c(Years, Salary) retains its column-selection meaning.

Chart(): Parameter facet now works for treemap, icicle charts, and bubble charts.

XY(): A multi-column data frame passed directly as x and y generates a scatterplot matrix, so the variable list can be stored once and reused: vec <- d[, c("Years","Pre","Post")] then XY(vec, vec, data=NULL).

XY(): Parameter 'facet' now works for all forms of relationship visualizations beyond only scatterplots.

XY(): With a vector of x (or y) variables, facet repeats the multi-series overlay across its panels, each panel the full overlay for one level of the facet variable, all panels on a common scale: XY(c(Pre, Post), Salary, facet=Gender). Parameter by does not apply with a vector of x or y variables, as the multiple series already occupy the color (series) dimension of the plot.

Regression() and Logit(): A term of the model formula may now be an expression of variables, such as reg(log(Salary) ~ Years + I(Years^2)), evaluated against the data frame with the expression text as the term label, so no need to first create the transformed variable with Transform().

ttest() and ANOVA(): Model terms may also be expressions, such as ttest(log(Salary) ~ Gender), ttest(log(Pre), log(Post)), or ANOVA(Salary ~ cut(Years, 3)), with the expression text as the label. An ANOVA() expression for a factor must produce categories, such as with cut() or factor(). Previously, ttest() with an expression response in a formula silently analyzed the untransformed variable.

Correlation(): The variables may also be expressions, such as Correlation(log(Years), Salary), with the expression text as the label. An expression that yields a data frame, such as Correlation(d[, 1:4]), specifies a correlation matrix.

Model(): Expanded to call more possibilities for analysis of a model.

Bug Fixes