rk.svyplot: Base Graphics for Survey Data

An RKWard plugin package that provides a graphical interface for the native plotting functions of the survey package. Unlike other plugins that use ggplot2, this package focuses on Base R graphics, offering specialized visualizations for complex survey designs (weighted data) that are lightweight and statistically rigorous.

It allows users to create survey-weighted plots—which are often more representative than standard plots—without writing R code.

What's New in Version 0.0.2

This version focuses on accessibility and internationalization. The entire plugin suite has been fully localized.

• Multilingual Support: The interface is now available in:
  • 🇺🇸 English (Default)
  • 🇪🇸 Spanish (es)
  • 🇫🇷 French (fr)
  • 🇩🇪 German (de)
  • 🇧🇷 Portuguese (Brazil) (pt_BR)

Features

The plugin provides a suite of tools organized under the Survey -> Graphs -> Survey plots menu in RKWard.

1. Survey Scatterplot (svyplot)

The main plugin. Creates survey-weighted scatterplots with several styles to visualize weights and density:

• Bubble: Point size is proportional to the survey weight.
• Hexagonal Bins: Uses hexagonal binning to show density (great for large datasets).
• Transparent: Uses alpha blending (transparency) to show overlapping points, with optional coloring by a third variable using hcl.colors.
• Subsample: Plots a random subsample of the data.

2. Bar Plot (barplot.svystat)

Creates a bar plot directly from a svytable object (produced by the rk.survey.design or rk.questionr plugins).

• Options: Side-by-side bars, legend toggles, and integrated ColorBrewer palettes.

3. Conditioned Scatterplot (svycoplot)

Creates faceted scatterplots (Trellis plots) to examine relationships conditioned on the levels of a third variable.

• Styles: Supports Hexagonal Bins and Transparency styles.
• Coloring: Options to control color schemes and transparency based on variable levels.

4. Survey Q-Q Plot (svyqqplot)

Produces quantile-quantile plots to compare the distribution of a survey variable against a theoretical distribution.

• Distributions: Compare against Normal, t, or Chi-squared distributions.
• Degrees of Freedom: Adjustable for t and Chi-squared tests.

5. Survey Histogram & Boxplot

A versatile component to visualize distributions.

• Histogram (svyhist): Weighted histogram with adjustable bin counts/breaks.
• Boxplot (svyboxplot): Weighted boxplots, optionally grouped by a factor variable.

6. Smoothed Scatterplot (svysmooth)

Creates a scatterplot with a smoothed conditional mean line (using LOESS or GAM) to visualize trends in survey data.

General Features

• NA Handling: Most scatterplot-based plugins include an option to automatically remove observations with missing values (NA) in the selected variables before plotting.
• Output Control: High-resolution export to PNG, JPG, and SVG with custom dimensions and resolution (PPI).
• Live Previews: All plugins include a Preview button, allowing you to see the plot and adjust settings before creating the final output.

Installation

1. Prerequisite: Ensure you have the remotes package installed in R.

2. Install: Run the following command in the RKWard R Console:

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

3. Restart: Restart RKWard to load the new menu entries and translations.

Usage Example: Creating a Conditioned Scatterplot

This example demonstrates how to create a faceted scatterplot to examine the relationship between two variables across different school types.

1. Prepare Sample Data First, load the survey library and create a svydesign object. Run this code in the RKWard console:

library(survey)
data(api)

# Create a survey design object
dclus1 <- svydesign(id=~dnum, weights=~pw, data=apiclus1, fpc=~fpc)

You will now have a survey design object named dclus1 in your workspace.

2. Open the Plugin Go to the RKWard menu: Survey -> Graphs -> Survey plots -> Conditioned Scatterplot.

3. Configure the Dialog

• Survey Design Object: From the object selector list on the left, choose dclus1.
• Y Variable: Select api00 (API score in 2000).
• X Variable: Select api99 (API score in 1999).
• Conditioning Variable (Facet): Select stype (School Type). This variable will create the panels.
• Navigate to the Options tab.
  • Ensure "Plot Style" is set to Hexagonal Bins.
  • (Optional) Adjust the palette settings.

4. Preview the Result Click the Preview button. A plot will appear in the dialog, showing the relationship between api00 and api99 in three separate panels: one for Elementary, one for Middle, and one for High Schools.

5. Submit Click the Submit button. A new plot device window will open in RKWard displaying the final, high-quality graph.

Dependencies

This plugin requires the following R packages:

• survey
• hexbin (for hexagonal binning styles)
• RColorBrewer
• KernSmooth (for smoothing)

Author & License

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