Glossary

Glossary

Core Objects

• Corpus: The unified analytical container holding textual Documents and a numeric DataFrame, along with visualization artifacts and metadata derived from analyses.
• Document: A textual item with id, name, text, and metadata (e.g., assigned topics/keywords, cluster labels).
• DataFrame (df): The numeric table (Pandas DataFrame) aligned to documents via a shared identifier (id) or through filters that propagate across modalities.
• Metadata: Key–value store attached to the corpus or documents holding results (e.g., decision_tree_feature_importance, regression_coefficients, pca, numeric_clusters).

Common CLI Flags

• --source/--sources: Import from folder(s) (reads .txt/.pdf and one .csv) or a URL.
• --inp / --out: Load/save corpus and analysis results.
• Textual analysis:
  • --topics, --assign, --summary, --sentiment, --nlp.
• Predictive / numeric analysis (requires [ml] extras):
  • --regression, --cart (Decision Trees), --kmeans, --pca, --cls (SVM), --nnet, --lstm.
• Visualization (separate CLI crispviz):
  • --wordcloud, --ldavis, --pca, --tdabm.
• Semantic and filtering (via crispt and crisp):
  • --filters key=value, --semantic, --semantic-chunks, --rec.

Typical Workflow

1. Import: crisp --source crisp_source --out crisp_input
2. Textual: crisp --inp crisp_input --topics --assign --sentiment --out crisp_input
3. Numeric: crisp --inp crisp_input --ml --regression --cart --kmeans --pca --out crisp_input
4. Visualize: crispviz --inp crisp_input --wordcloud --ldavis --pca --out viz_out/
5. Triangulate: crisp --inp crisp_input --print metadata --print regression_coefficients

Inline Notes

• Topic number defaults to 8 for LDA visualization guidance and can be changed with --num/--topics-num (Mettler et al., 2025).
• TDABM supports shape-aware exploration of multi-dimensional numeric data (Rudkin & Dlotko, 2024).

See Also

• Getting Started
• Predictive Methods
• Topological Data Analysis
• Human-Sense-making-and-triangulation