rk.correspondence: Multivariate Analysis for RKWard

rk.correspondence brings modern, visually appealing multivariate statistics to the RKWard GUI. Powered by the industry-standard FactoMineR and factoextra packages, this plugin provides an intuitive, "R Commander-style" interface to compute and visualize Simple and Multiple Correspondence Analysis without writing a single line of code.

🚀 What's New in Version 0.0.1

This is the initial release of the package, introducing two core statistical tools for categorical data exploration:

1. Simple Correspondence Analysis (CA): Analyze two-way contingency tables, customize biplots, and filter categories by their statistical contribution or quality of representation ($cos^2$).
2. Multiple Correspondence Analysis (MCA): Explore patterns across multiple categorical variables simultaneously, isolate individuals or variables in the plot, and control dimensionality.

🌍 Internationalization

The interface is fully localized in:

• 🇺🇸 English (Default)
• 🇪🇸 Spanish (es)
• 🇫🇷 French (fr)
• 🇩🇪 German (de)
• 🇧🇷 Portuguese (Brazil) (pt_BR)

✨ Features

1. Simple CA (Contingency Tables)

• Dimensionality Control: Choose how many dimensions to retain (ncp) and exactly which axes to plot (e.g., Dim 1 vs Dim 3).
• Advanced Filtering: Declutter your plots by showing only the "Top N Contributors" or points with a specific quality of representation ($cos^2$) threshold.
• Aesthetics: Instantly change row and column colors, and use repel to prevent text labels from overlapping.

2. Multiple CA (Large Datasets)

• Active Variable Selection: Easily select multiple categorical variables from a large data frame using a multi-selection list.
• Plot Isolation: Toggle between plotting the full Biplot, Variables only, or Individuals only to uncover different data structures.

🛡️ Universal Features

• Live Plot Preview: Adjust colors, axes, and filters while seeing the ggplot2-based biplot update in real-time before running the final model.
• Object Saving: Automatically saves the complex FactoMineR result objects to your R Global Environment for further scripting or exporting.

📦 Installation

This plugin is available via GitHub. Use the remotes or devtools package in RKWard to install it.

1. Open RKWard.

2. Run the following command in the R Console:

   # If you don't have devtools installed:
   # install.packages("devtools")
   
   local({
     require(devtools)
     install_github("AlfCano/rk.correspondence", force = TRUE)
   })
   

3. Restart RKWard to load the new menu entries.

💻 Usage & Testing Instructions

Once installed, the tools are organized under: Analysis -> Multivariate

You can test the plugins immediately using R's built-in datasets. Copy and paste this into your RKWard console to prepare the test data:

# 1. Data for Simple CA (Contingency table)
tabla_ca <- margin.table(HairEyeColor, c(1, 2))

# 2. Data for Multiple CA (Dataset 'tea' from FactoMineR)
data(tea, package = "FactoMineR")
datos_mca <- tea[, c("Tea", "How", "how", "sugar", "where", "always")]

Test 1: Simple CA

1. Go to Analysis -> Multivariate -> Simple CA.
2. Data tab: Select tabla_ca.
3. Plot Options tab: Click Preview. Change the "Row color" to darkgreen and "Column color" to purple. Try filtering by "Top Contributors" and set the threshold to 3.
4. Click Submit. Check your Output window for the statistical summary and the final plot.

Test 2: Multiple CA

1. Go to Analysis -> Multivariate -> Multiple CA.
2. Data & Variables tab: Select datos_mca as the Dataset. Select all variables in the list and move them to "Active Categorical Variables".
3. Plot Options tab: Click Preview. Change "Plot Type" to Variables only to see a cleaner graph.
4. Click Submit. Check your R Workspace; you will see the resultado_mca object saved and ready to use.

🛠️ Dependencies

This plugin relies on the following R packages:

• FactoMineR (Statistical computation)
• factoextra (ggplot2-based visualization)
• rkwarddev (Plugin generation)

✍️ Author & License

• Author: Alfonso Cano (alfonso.cano@correo.buap.mx)
• Assisted by: Gemini, a large language model from Google.
• License: GPL (>= 3)