PyCaret
GitHub
Module

pycaret.plots.eda

Plotly-native exploratory data analysis (EDA) plots.

Functions:

  • column_distribution — histogram (numeric) / bar (categorical) for a single column.
  • correlation_heatmap — Pearson correlation matrix as a heatmap.
  • missingness_map — per-column missing-rate bar.
  • target_vs_feature — joint distribution of a feature and the target (boxplot for numeric→categorical, scatter for numeric→numeric, count-bar for categorical→categorical).
  • profile_summary — per-column data-quality summary table figure.

Functions 5

column_distribution(data: Any, column: str, bins: int = 30, title: str | None = None) -> go.Figure

Histogram (numeric) or count-bar (categorical) for one column.

correlation_heatmap(data: Any, method: str = 'pearson', annotate: bool = True, title: str | None = 'Correlation matrix') -> go.Figure

Numeric correlation matrix as a heatmap.

method: pearson (default), spearman, kendall.

missingness_map(data: Any, title: str | None = 'Missing data per column') -> go.Figure

Per-column missing-rate horizontal bar chart.

target_vs_feature(data: Any, feature: str, target: str, title: str | None = None) -> go.Figure

Joint plot of feature × target.

  • numeric × numeric → scatter with trend line
  • numeric × categorical → boxplot grouped by target
  • categorical × numeric → boxplot grouped by feature
  • categorical × categorical → grouped bar (count)
profile_summary(data: Any, title: str | None = 'Column profile') -> go.Figure

Per-column data-quality summary as a Plotly table figure.

Columns: Name, Type, Non-null %, Unique, Mean, Std, Min, Max. Useful as the EDA screen's "headline" panel.