clophfit.fitting.diagnostic_plots#

Reusable matplotlib-only diagnostic plots for standardized residuals.

These plots are independent of seaborn and are suitable for both package tests and manuscript workflows.

Functions#

plot_residual_overview(residuals, *[, residual_col, ...])

Create a generic standardized-residual diagnostic overview.

Module Contents#

clophfit.fitting.diagnostic_plots.plot_residual_overview(residuals, *, residual_col='std_res', x_col='x', label_col='label', well_col='well', step_col='step', output_path=None, title=None, bins=40, alpha=0.45)#

Create a generic standardized-residual diagnostic overview.

Panels: (A) distribution, (B) residuals vs x, (C) lag-1 correlation histograms, (D) normal Q-Q plot. Uses ±2 visual guides.

The residual column should contain model-standardized residuals:

std_res = (observed - predicted) / sigma_used_in_fit

not a global z-score of raw residuals.

Parameters:

  • residuals (pandas.DataFrame)

  • residual_col (str)

  • x_col (str)

  • label_col (str)

  • well_col (str)

  • step_col (str)

  • output_path (str | pathlib.Path | None)

  • title (str | None)

  • bins (int)

  • alpha (float)

Return type:

matplotlib.figure.Figure