Skip to content

Visualization Commands

Jump to bottom
Bell Eapen edited this page Jan 29, 2026 · 1 revision

Visualization Commands

The crispviz tool creates publication-quality visualizations from your analyzed corpus. All commands operate on a corpus that has already been processed with crisp or crispt.

General Syntax:

crispviz --inp <corpus_path> --out <viz_output_dir> [flags]

Text Visualizations

Word Cloud

Generate a word cloud for each identified topic.

  • Command: crispviz --wordcloud
  • Prerequisite: Run crisp --topics first.

Word Frequency

Plot the frequency distribution of words in the corpus.

  • Command: crispviz --freq
  • Options: --bins <N> (default: 100).

Top Terms

Bar chart of the most frequent terms.

  • Command: crispviz --top-terms
  • Options: --top-n <N> (default: 20).

Interactive LDA (PyLDAvis)

Create an interactive HTML visualization of your topic model.

  • Command: crispviz --ldavis
  • Prerequisite: Run crisp --topics first.
  • Output: Saves an .html file instead of an image.

Statistical Visualizations

Correlation Heatmap

Visualize correlations between numeric columns in your dataframe.

  • Command: crispviz --corr-heatmap
  • Options: --corr-columns "col1,col2" (Comma-separated list of columns to include. Auto-selects if empty).

Structural & Network Visualizations

Knowledge Graph

Visualize the network of documents, keywords, and metadata.

  • Command: crispviz --graph
  • Prerequisite: Run crisp --topics (for keywords) or crispt --graph.
  • Options:
    • --graph-nodes <types>: Filter nodes (e.g., document,keyword). Default: all.
    • --graph-layout <algo>: Layout algorithm (spring, circular, kamada_kawai, spectral).

TDA Network (TDABM)

Visualize the Topological Data Analysis network.

  • Command: crispviz --tdabm
  • Prerequisite: Run crispt --tdabm ... first.

Topic Distribution

Plot the distribution of dominant topics across documents.

  • Command: crispviz --by-topic
  • Prerequisite: Run crisp --assign first.

Common Options

  • --out <dir>: Required. Directory to save the PNG/HTML files.
  • --inp <dir>: Input corpus directory.
  • --topics-num <N>: Number of topics to assume if regenerating models (default: 8).

Clone this wiki locally