Major New Features

Any features added to lessR since August, 2013 are not presented in the book. Enter Help(lessR), then click on Package NEWS to see the full set of additions and bug fixes.

• Data frame subset. New function .() for general, easy subsetting of a data frame using the
  d[.(rows),.(cols)]
  format for a data frame named d.
• Pivot table. Aggregate data with a new function pivot() for creating a 1-D pivot table for input into subsequent analysis/visualization functions or as a 2-D table for display.
• pad_x and pad_y parameters. Added to provide optional additional padding beyond the default 0 padding of the x- and y-axes.
• see(). New function to view the top-left and bottom-right corners of a data frame, especially appropriate for large data frames.
• Read SPSS data files. The haven package is now used instead of foreign. Now value labels and variable labels are recovered. Each SPPS integer categorical variable with value labels now imported as an integer variable and also a factor with the value labels.
• Multiple functions. Default data frame name is now d, instead of mydata, though the later still works.
• Multiple functions. Parameter names now delimited with an underline instead of a period, as in bin_width instead of bin_width. Old versions still work.
• Multiple functions. Parameter filter added to subset rows directly from the function call. Provide a logical condition such as (Gender=="M").
• getColors(). Function provides many options for generating color palettes. See the examples in the help function for examples. Then use these palettes for parameter fill for the visualization functions.
• Regression(). K-fold cross-validation added with the k-fold parameter. Set to an integer for the number of folds.
• Logit(). Probability threshold parameter prob_cut added so default of 0.5 can be changed for predicting (classifying) group membership. In addition to accuracy, recall and precision fit indices added, computed from the confusion matrix.
• Histogram(). Now can get the density plot overlaid on the histogram with the Histogram function with density=TRUE.
• Plot(). Add many features such as the 0.95 confidence ellipse and outlier identification, best-fitting line with and without outliers with enhance=TRUE.
• Plot(). For any numeric variable, result is the newly developed VBS Plot, an integrated Violin/Box/Scatterplot, with the output tuned according to sample size and distribution characteristics. Replaces the pre-computer technology of the histogram with the modern alternative for the display of the distribution of a continuous variable. Trellis plots also available with the by1 and by2 parameters, plus the scatterplot can be shown more multiple groups on a single panel with the by parameter.
• Plot(). New parameter enhance=TRUE adds much information to the 2-variable scatterplot.
• Plot(). MD.cut=0, out.cut=0 added to label outliers in box plot (part of the VBS plot) and two-variable scatterplots.
• Plot(), Histogram(), Bar Chart(). Trellis plots added with the by1 and by2 parameters.
• Plot(). Option topic added for analysis of means, counts, etc. instead of just the original data, and option object added to specify objects other than points.
• BarChart(). y parameter added to allow a bar chart of any y-variable with new parameter order -- x, y, by -- so by must now be explicitly indicated to specify a by variable
• Plot(). Revised function plots a scatter plot with any combination of one or more continuous (numeric) or categorical (non-numeric) variables.
• Read(). Parameter var_labels=TRUE to indicate a file of variable labels. Read into the data frame named l.
• Automatic R Markdown. The knitr compatibility incentivized a new approach to output in which the output is now generated in segments, currently implemented for some functions, such as Regression, ANOVA, Histogram, Density, and ScatterPlot (1-variable). The segments are automatically collated for the output, or, can be stored in an object such as

  r <- reg(Y ~ X)

  To see the names of all the segments, enter

  names(r)

  referring to the name of the output object, here r. Then the segments can be listed separately, such as

  r$out_estimates

  In addition, the

  Rmd

  option generates a specified R markdown file that can be run with additional interpretative text. This option is currently available for Regression, Histogram (and BoxPlot, Density and ScatterPlot) and cfa functions.
  
  The interpretative output for Regression is well along in development. For example, the following code generates the R markdown file reg.Rmd.

  d <- rd("Reading")
  reg(Reading ~ Verbal + Absent + Income, Rmd="reg")
  

  The resulting markdown file is automatically converted to a web page in RStudio.
• Read Excel data files. The lessR function Read can now directly read Excel data files. This functionality is from Hadley Wickham's readxl package. Both the main data file and/or the labels files can now be Excel files.
• Browse for the variable labels file. The labels files in the Read statement can now be specified as labels="", which allows the user to browse for the labels file.
• Variable labels added/modified directly. Variable labels can now be entered or modified directly with the label function in addition to the option of reading directly from a file.
• Reduced Read() output option. A brief version of the details function is now available for providing information about a data frame, called details.brief. This function is also called by Read and is invoked with the abbreviation rd.brief.
• Analysis of multiple variables extended to Density and LineChart. An entire data frame or a list of variables defined as with the c function can now be specified for analysis with the Density and LineChart functions in addition to BarChart, BoxPlot, and Histogram.
• Dependent-groups t-test plot. The ttest function, when applied to dependent groups with paired=TRUE, now provides a plot of each pair of data values in addition to the density plot of differences. This new plot includes a diagonal line that represents the equality of both values, and a line segment for each point to indicate the vertical distance from this line.