MMMPlotSuite.posterior_predictive#

MMMPlotSuite.posterior_predictive(var=None, idata=None, hdi_prob=0.85, backend=None)[source]#

Plot posterior predictive distributions over time.

Visualizes posterior predictive samples as time series, showing the median line and highest density interval (HDI) bands. Useful for checking model fit and understanding prediction uncertainty.

Parameters:

varstr, optional

Variable name to plot from posterior_predictive group. If None, uses “y”.

idataxr.Dataset, optional

Dataset containing posterior predictive samples with a “date” coordinate. If None, uses self.idata.posterior_predictive.

This parameter allows: - Testing with mock data without modifying self.idata - Plotting external posterior predictive samples - Comparing different model fits side-by-side

hdi_probfloat, default 0.85

Probability mass for HDI interval (between 0 and 1).

backendstr, optional

Plotting backend to use. Options: “matplotlib”, “plotly”, “bokeh”. If None, uses global config via mmm_plot_config[“plot.backend”]. Default is “matplotlib”.

Returns:

PlotCollection

arviz_plots PlotCollection object containing the plot.

Use .show() to display or .save("filename") to save. Unlike the legacy suite which returned (Figure, Axes), this provides a unified interface across all backends.

Raises:

ValueError: If no posterior_predictive data found in self.idata and no idata provided.
ValueError: If hdi_prob is not between 0 and 1.

See also

LegacyMMMPlotSuite.posterior_predictive: Legacy matplotlib-only implementation

Notes

Breaking changes from legacy implementation:

Returns PlotCollection instead of (Figure, Axes)
Different interface for saving and displaying plots

Examples

Basic usage:

mmm.sample_posterior_predictive(X)
pc = mmm.plot.posterior_predictive()
pc.show()

Plot with different HDI probability:

pc = mmm.plot.posterior_predictive(hdi_prob=0.94)
pc.show()

Save to file:

pc = mmm.plot.posterior_predictive()
pc.save("posterior_predictive.png")

Use different backend:

pc = mmm.plot.posterior_predictive(backend="plotly")
pc.show()

Provide explicit data:

external_pp = xr.Dataset(...)  # Custom posterior predictive
pc = mmm.plot.posterior_predictive(idata=external_pp)
pc.show()

Direct instantiation pattern:

from pymc_marketing.mmm.plot import MMMPlotSuite

mps = MMMPlotSuite(custom_idata)
pc = mps.posterior_predictive()
pc.show()