SDAnalysis-kun
src/sdanalysis_kun/ contains a Tkinter-based GUI application. It is a companion tool for trying the analyses from the lesson code through CSV selection and button-based operations.
Downloads
Launch command:
uv run sdanalysis-kun
Alternatively:
uv run sd-method-lessons
The GUI mainly supports the following operations:
- Select a CSV file
- Select the stimulus column
- Select the respondent column (optional)
- Filter the stimuli to analyze
- Select adjective-pair columns
- Specify a regular expression for formatting adjective-pair names
- Select a 5-point or 7-point rating scale
- Run parallel analysis with Pearson correlation or polychoric correlation
- Specify the number of factors, or use the number suggested by the parallel analysis (PA) result
- Choose Promax rotation, Varimax rotation, or no rotation
- Display, plot, and export the factor loading matrix as CSV
- Display and export the mean and standard deviation (SD) of factor scores by stimulus as CSV
- Display a stimulus map with PCA based on mean factor scores; when a respondent column is selected, each stimulus is summarized by its centroid and a within-stimulus 1-SD covariance ellipse representing respondent variability
Expected CSV Structure
SDAnalysis-kun assumes a table with the following layout.
- The data is in long format for observations: each row is one response record (typically one respondent rating one stimulus).
- A
stimulus_idcolumn exists and stores the stimulus identifier. - A
respondent_idcolumn exists and stores the respondent identifier. - Adjective-pair ratings are in wide format: each adjective pair is a separate numeric column.
Example:
id,respondent_id,stimulus_id,fast-slow,bright-dark,soft-hard
1,r_001,s_001,5,2,4
2,r_001,s_002,3,4,2
3,r_002,s_001,6,3,5
In short: observation keys (stimulus_id, respondent_id) are row-wise, and adjective pairs are column-wise.