Choosing a plot

surveyframe draws in 3 ways, and which one fits depends on what you already hold.

From a report or a set of results

Every report that has a natural chart carries a plot() method, so a report draws without naming a helper:

• plot(reliability_report(...)) draws one report.

• plot(results) draws each block of an analysis that has a chart.

• plot(results, which = "RQ1") selects one block by its ID.

This is the shortest route, and it is the one to reach for first.

From a helper, by what it takes

The helpers exist for assembling a custom report, where you need one chart on its own terms. They differ in what they accept and in what comes back.

• A report object, giving 1 plot:

  • sframe_plot_reliability()

  • sframe_plot_validity()

  • sframe_plot_quality()

  • sframe_plot_missingness()

  • sframe_plot_efa_scree()

  • sframe_plot_efa_loadings()

• Responses with the instrument, giving 1 plot:

  • sframe_plot_item_chart()

  • sframe_plot_scale_chart()

  • sframe_plot_likert_matrix()

  • sframe_plot_likert_scale()

  • sframe_plot_correlation_matrix()

  • sframe_plot_descriptives()

• One block's result, giving 1 plot:

  • sframe_plot_group_comparison()

  • sframe_plot_paired_comparison()

  • sframe_plot_decision_ranking()

  • sframe_plot_dematel_influence()

• A regression result, giving 4 panels as a list:

  • sframe_plot_regression_diagnostics()

• Responses and 1 column, giving 3 panels as a list:

  • sframe_plot_variable_distribution()

• A text result, giving 1 plot:

  • sframe_plot_term_frequency()

  • sframe_plot_ngram_frequency()

  • sframe_plot_cooccurrence()

  • sframe_plot_cooccurrence_network()

  • sframe_plot_sentiment()

  • sframe_plot_topics()

A helper that finds nothing to draw returns NULL, so guard the result where a report has to keep rendering.

Integration helpers

sframe_draw_mosaic() and sframe_draw_likert_diverging() draw into an open device for the generated reports, and stay exported so those reports keep working. sframe_likert_scale_groups() finds the scales whose items share a choice set, which is how a report groups them. Take the results they draw from run_analysis_plan().

See also

run_analysis_plan(), render_report()