BREAKING
gg_rsm() arguments filled, size, stroke have been removed. They were used for metR::geom_text_contour which is no longer being used.
pair_plots no longer checks design is a proper Latin hypercube design
NEW FEATURES
gg_rsm() has new argument bins to specify the number of contour bins. Default is 6gg_rsm() has new argument text_size to specify size of text for labelled contour lines. Default is 3design_to_tibble() to convert an object of class design to tibble. Designs generated from packages such as DoE.base, FrF2BUG FIXES
main_effects now has independent x-axis limits for each factor. Input design properly accepts class “design” generated from packages such as DoE.base, FrF2, same goes for interaction_effectsn_columns arguments is automatically adjusted to n_columns=1 in main_effects,interaction_effects, pair_plots whenever there is only one factor or pair of factorsDEPENDENCIES
metR to geomtextpath. geomtextpath is a lighter package in terms of dependencies compared to metR. This only affects gg_rsm and is used to add text labels on contour linesBREAKING
diagnostic_plots() to gg_lm()gg_boxplots() arguments factor,response to x,y, respectively.twoD_projections() to pair_plots()form argument to formula in gg_rsm()NEW FEATURES
pulp_experiment data from “Experiments: Planning, Analysis, and Optimization, CFJ Wu, MS Hamada - Second edition”girder_experimet data from “Experiments: Planning, Analysis, and Optimization, CFJ Wu, MS Hamada - Second edition”effect_colors to pareto_plot(). Default are c('#d9a698','#9ecede') for positive and negative effects, respectively.DEPENDENCIES
data.table as a dependency. Now there are only two imports ggplot2 and insightgridExtra to patchwork. No functionality change.BUG FIXES
gg_rsm where numbers were not rounded and displayed over 10 digits. This resulted in labels in figures being cut off or stacked on top of each otherBREAKING
gg_boxplots() arguments response,factor,group_var are now string charactersNEW FUNCTIONS
theme_bw_nogrid(), simple theme settings which were used throughout ggDoE plotsCHANGES
removed dplyr dependency in favor of data.table. data.table has significantly less dependencies compared to dplyr.
added insightto Imports for checking if required packages are installed for each respective function. insight will later be used for cleaning up source code for extracting information from model objects such as variables, responses, etc…
switched gridExtra and unrepx from Imports to Suggests
MINOR CHANGES
BREAKING
ncols to n_columns in all functions which previously used this argument. These functions are: diagnostic_plots(),gg_rsm(), interaction_effects(),main_effects(),twoD_projections()NEW FEATURES
half_normal() has new argument point_color to change the color of the pointsdiagnostic_plots() has a new argument cooksD_type to change the threshold computed for Cook’s distance plot (plot #6)CHANGES
half_normal() now uses ggrepel::geom_text_repel instead of hard-coded geom_text to try and avoid label overlapunrepex from Suggests to ImportsNEW FEATURES
changed appearance of gg_boxplots() from using fill to color. Added new argument jitter_points to gg_boxplots() to overlay points to boxplots.
Added initial vignette for a quick overview of ggDoE. Still needs further details for certain plots
BUX FIXES
twoD_projections() that would output the same plot for every plot in the gridNEW FUNCTIONS
twoD_projections(). This function will output all two dimensional projections from a Latin hypercube design. Will add further details to documentation in future updates.CHANGES
gridExtra to Imports instead of Suggests under Description fileCHANGES
diagnostic_plots() function. This is the sixth diagnostic plot.BREAKING
gg_rsm() restricted model class to only be an rsm object. Future updates will add back lm class object option.BUG FIXES
diagnostic_plots() to check if there is insufficient degrees of freedom to obtain standard errors, t.value. It won’t plot any of the diagnostic plots for that model.MINOR CHANGES
NEWS.md file to track changes to the package.CURRENT FUNCTIONS
alias_matrix() Color Map on Correlations
boxcox_transform() Box-Cox transformation plot with 95% confidence interval of lambda values to consider
gg_boxplots() Boxplots created with ggplot2
diagnostic_plots() Regression Diagnostic Plots with ggplot2
gg_rsm() Contour plot(s) of a fitted linear model in ggplot2 from an RSM object
half_normal() A halfnormal plot for factorial effects
lambda_plot() Lambda plot for tracing t-statistics across different values of lambda
main_effects() Main effects plots for each factor in a design given a specified response
pareto_plot() A bar plot with ordered effects, margin of error (ME) and simultaneous margin of error (SME) cutoffs.
viridisPalette() A simple wrapper for the viridisLite package used for changing the color palette in most plots above