Method Selection Guide

Method Selection Guide

This table helps newcomers choose the appropriate CRISP-T analysis method based on their data type and research goal.

Goal / Data Type	Textual Data (documents)	Numeric Data (DataFrame)	Both (Mixed)	Recommended CLI Flags / Tools
Topic discovery	Yes	No	Yes (text)	--topics, --assign, crispviz --ldavis
Sentiment analysis	Yes	No	Yes (text)	--sentiment, --sentence
Summarization	Yes	No	Yes (text)	--summary
Clustering	No	Yes	Yes	--ml --kmeans, --pca, crispviz --pca
Predictive modeling (outcome)	No	Yes	Yes	--ml --cart (Decision Tree), --regression
Topological structure (TDABM)	No	Yes	Yes	crispt --tdabm, crispviz --tdabm
Semantic search/filtering	Yes	No	Yes (text)	--semantic, --filters key=value
Triangulation (mixed methods)	Yes	Yes	Yes	Combine above; inspect metadata, use --print
Visualization	Yes	Yes	Yes	crispviz --wordcloud --ldavis --pca --tdabm

Notes:

• For mixed data, always import both text and numeric files to the same corpus.
• Use --ml for machine learning features (requires [ml] install).
• See Glossary for flag definitions and Examples for full workflows.

See Also

• Getting Started
• Examples
• Glossary