plot_distributions

Distribution plotting module for ML results analysis. Focuses on metric distributions with outcome variable stratification.

Attributes

MAX_OUTCOMES_FOR_STRATIFIED_PLOT

MAX_OUTCOMES_FOR_HEATMAP

Classes

DistributionPlotter

Initializes the DistributionPlotter.

Functions

plot_metric_correlation_by_outcome(data[, ...])

Plots correlation matrices of metrics, stratified by outcome variable.

Module Contents

plot_distributions.MAX_OUTCOMES_FOR_STRATIFIED_PLOT = 20[source]
plot_distributions.MAX_OUTCOMES_FOR_HEATMAP = 25[source]
class plot_distributions.DistributionPlotter(data: pandas.DataFrame, style: str = 'default')[source]

Initializes the DistributionPlotter.

Parameters:

  • data (pd.DataFrame) – A DataFrame containing the experiment results.

  • style (str, optional) – The matplotlib style to use for plots. Defaults to ‘default’.

data[source]
clean_data[source]
colors[source]
plot_metric_distributions(metrics: List[str] | None = None, figsize: Tuple[int, int] = (15, 10), stratify_by_outcome: bool = False, outcomes_to_plot: List[str] | None = None)[source]

Plots distributions of key performance metrics.

Parameters:

  • metrics (Optional[List[str]], optional) – A list of metric columns to plot. If None, uses a default list. Defaults to None.

  • figsize (Tuple[int, int], optional) – The figure size. Defaults to (15, 10).

  • stratify_by_outcome (bool, optional) – If True, creates separate plots for each outcome. Defaults to False.

  • outcomes_to_plot (Optional[List[str]], optional) – A list of specific outcomes to plot. If None, all are used. Defaults to None.

Raises:

ValueError – If no specified metrics are found in the data.

plot_comparative_distributions(metric: str = 'auc', outcomes_to_compare: List[str] | None = None, plot_type: str = 'overlay', figsize: Tuple[int, int] = (12, 6))[source]

Creates comparative distribution plots across different outcomes.

Parameters:

  • metric (str, optional) – The metric to compare. Defaults to ‘auc’.

  • outcomes_to_compare (Optional[List[str]], optional) – A list of specific outcomes to compare. If None, all are used. Defaults to None.

  • plot_type (str, optional) – The type of plot to generate: ‘overlay’, ‘subplot’, or ‘violin’. Defaults to ‘overlay’.

  • figsize (Tuple[int, int], optional) – The figure size. Defaults to (12, 6).

Raises:

ValueError – If ‘outcome_variable’ or the specified metric is not found, or if an invalid plot_type is provided.

plot_distribution_summary_table(metrics: List[str] = None, outcomes_to_include: List[str] | None = None) → pandas.DataFrame[source]

Creates a summary table of distribution statistics by outcome.

Parameters:

  • metrics (Optional[List[str]], optional) – A list of metrics to include in the summary. If None, uses a default list. Defaults to None.

  • outcomes_to_include (Optional[List[str]], optional) – A list of specific outcomes to include. If None, all are used. Defaults to None.

Returns:

A DataFrame containing the summary statistics.

Return type:

pd.DataFrame

Raises:

ValueError – If ‘outcome_variable’ column is not found.

plot_distribution_heatmap(metrics: List[str] = None, stat: str = 'mean', figsize: Tuple[int, int] = (10, 6))[source]

Creates a heatmap of distribution statistics across outcomes and metrics.

Parameters:

  • metrics (Optional[List[str]], optional) – A list of metrics to include. If None, uses a default list. Defaults to None.

  • stat (str, optional) – The statistic to display (‘mean’, ‘std’, ‘median’, ‘min’, ‘max’). Defaults to ‘mean’.

  • figsize (Tuple[int, int], optional) – The figure size. Defaults to (10, 6).

Raises:

ValueError – If ‘outcome_variable’ column is not found.

plot_distributions.plot_metric_correlation_by_outcome(data: pandas.DataFrame, outcomes_to_plot: List[str] | None = None, figsize: Tuple[int, int] = (15, 10))[source]

Plots correlation matrices of metrics, stratified by outcome variable.

Parameters:

  • data (pd.DataFrame) – The results DataFrame.

  • outcomes_to_plot (Optional[List[str]], optional) – A list of specific outcomes to plot. If None, all are used. Defaults to None.

  • figsize (Tuple[int, int], optional) – The figure size. Defaults to (15, 10).

Raises:

ValueError – If ‘outcome_variable’ column is not found.