Source code for arviz.plots.ppcplot

"""Posterior/Prior predictive plot."""

import logging
import warnings
from numbers import Integral

import numpy as np

from ..labels import BaseLabeller
from ..sel_utils import xarray_var_iter
from ..rcparams import rcParams
from ..utils import _var_names
from .plot_utils import default_grid, filter_plotters_list, get_plotting_function

_log = logging.getLogger(__name__)

[docs] def plot_ppc( data, kind="kde", alpha=None, mean=True, observed=None, observed_rug=False, color=None, colors=None, grid=None, figsize=None, textsize=None, data_pairs=None, var_names=None, filter_vars=None, coords=None, flatten=None, flatten_pp=None, num_pp_samples=None, random_seed=None, jitter=None, animated=False, animation_kwargs=None, legend=True, labeller=None, ax=None, backend=None, backend_kwargs=None, group="posterior", show=None, ): """ Plot for posterior/prior predictive checks. Parameters ---------- data : InferenceData :class:`arviz.InferenceData` object containing the observed and posterior/prior predictive data. kind : str, default "kde" Type of plot to display ("kde", "cumulative", or "scatter"). alpha : float, optional Opacity of posterior/prior predictive density curves. Defaults to 0.2 for ``kind = kde`` and cumulative, for scatter defaults to 0.7. mean : bool, default True Whether or not to plot the mean posterior/prior predictive distribution. observed : bool, optional Whether or not to plot the observed data. Defaults to True for ``group = posterior`` and False for ``group = prior``. observed_rug : bool, default False Whether or not to plot a rug plot for the observed data. Only valid if `observed` is `True` and for kind `kde` or `cumulative`. color : list, optional List with valid matplotlib colors corresponding to the posterior/prior predictive distribution, observed data and mean of the posterior/prior predictive distribution. Defaults to ["C0", "k", "C1"]. grid : tuple, optional Number of rows and columns. Defaults to None, the rows and columns are automatically inferred. figsize : tuple, optional Figure size. If None, it will be defined automatically. textsize : float, optional Text size scaling factor for labels, titles and lines. If None, it will be autoscaled based on ``figsize``. data_pairs : dict, optional Dictionary containing relations between observed data and posterior/prior predictive data. Dictionary structure: - key = data var_name - value = posterior/prior predictive var_name For example, ``data_pairs = {'y' : 'y_hat'}`` If None, it will assume that the observed data and the posterior/prior predictive data have the same variable name. var_names : list of str, optional Variables to be plotted, if `None` all variable are plotted. Prefix the variables by ``~`` when you want to exclude them from the plot. filter_vars : {None, "like", "regex"}, default None If `None` (default), interpret var_names as the real variables names. If "like", interpret var_names as substrings of the real variables names. If "regex", interpret var_names as regular expressions on the real variables names. A la ``pandas.filter``. coords : dict, optional Dictionary mapping dimensions to selected coordinates to be plotted. Dimensions without a mapping specified will include all coordinates for that dimension. Defaults to including all coordinates for all dimensions if None. flatten : list List of dimensions to flatten in ``observed_data``. Only flattens across the coordinates specified in the ``coords`` argument. Defaults to flattening all of the dimensions. flatten_pp : list List of dimensions to flatten in posterior_predictive/prior_predictive. Only flattens across the coordinates specified in the ``coords`` argument. Defaults to flattening all of the dimensions. Dimensions should match flatten excluding dimensions for ``data_pairs`` parameters. If ``flatten`` is defined and ``flatten_pp`` is None, then ``flatten_pp = flatten``. num_pp_samples : int The number of posterior/prior predictive samples to plot. For ``kind`` = 'scatter' and ``animation = False`` if defaults to a maximum of 5 samples and will set jitter to 0.7. unless defined. Otherwise it defaults to all provided samples. random_seed : int Random number generator seed passed to ``numpy.random.seed`` to allow reproducibility of the plot. By default, no seed will be provided and the plot will change each call if a random sample is specified by ``num_pp_samples``. jitter : float, default 0 If ``kind`` is "scatter", jitter will add random uniform noise to the height of the ppc samples and observed data. animated : bool, default False Create an animation of one posterior/prior predictive sample per frame. Only works with matploblib backend. To run animations inside a notebook you have to use the `nbAgg` matplotlib's backend. Try with `%matplotlib notebook` or `%matplotlib nbAgg`. You can switch back to the default matplotlib's backend with `%matplotlib inline` or `%matplotlib auto`. If switching back and forth between matplotlib's backend, you may need to run twice the cell with the animation. If you experience problems rendering the animation try setting ``animation_kwargs({'blit':False})`` or changing the matplotlib's backend (e.g. to TkAgg) If you run the animation from a script write ``ax, ani = az.plot_ppc(.)`` animation_kwargs : dict Keywords passed to :class:`matplotlib.animation.FuncAnimation`. Ignored with matplotlib backend. legend : bool, default True Add legend to figure. labeller : labeller, optional Class providing the method ``make_pp_label`` to generate the labels in the plot titles. Read the :ref:`label_guide` for more details and usage examples. ax : numpy array-like of matplotlib_axes or bokeh figures, optional A 2D array of locations into which to plot the densities. If not supplied, Arviz will create its own array of plot areas (and return it). backend : str, optional Select plotting backend {"matplotlib","bokeh"}. Default to "matplotlib". backend_kwargs : dict, optional These are kwargs specific to the backend being used, passed to :func:`matplotlib.pyplot.subplots` or :func:`bokeh.plotting.figure`. For additional documentation check the plotting method of the backend. group : {"prior", "posterior"}, optional Specifies which InferenceData group should be plotted. Defaults to 'posterior'. Other value can be 'prior'. show : bool, optional Call backend show function. Returns ------- axes : matplotlib_axes or bokeh_figures ani : matplotlib.animation.FuncAnimation, optional Only provided if `animated` is ``True``. See Also -------- plot_bpv : Plot Bayesian p-value for observed data and Posterior/Prior predictive. plot_loo_pit : Plot for posterior predictive checks using cross validation. plot_lm : Posterior predictive and mean plots for regression-like data. plot_ts : Plot timeseries data. Examples -------- Plot the observed data KDE overlaid on posterior predictive KDEs. .. plot:: :context: close-figs >>> import arviz as az >>> data = az.load_arviz_data('radon') >>> az.plot_ppc(data, data_pairs={"y":"y"}) Plot the overlay with empirical CDFs. .. plot:: :context: close-figs >>> az.plot_ppc(data, kind='cumulative') Use the ``coords`` and ``flatten`` parameters to plot selected variable dimensions across multiple plots. We will now modify the dimension ``obs_id`` to contain indicate the name of the county where the measure was taken. The change has to be done on both ``posterior_predictive`` and ``observed_data`` groups, which is why we will use :meth:`` to apply the same function to both groups. Afterwards, we will select the counties to be plotted with the ``coords`` arg. .. plot:: :context: close-figs >>> obs_county = data.posterior["County"][data.constant_data["county_idx"]] >>> data = data.assign_coords(obs_id=obs_county, groups="observed_vars") >>> az.plot_ppc(data, coords={'obs_id': ['ANOKA', 'BELTRAMI']}, flatten=[]) Plot the overlay using a stacked scatter plot that is particularly useful when the sample sizes are small. .. plot:: :context: close-figs >>> az.plot_ppc(data, kind='scatter', flatten=[], >>> coords={'obs_id': ['AITKIN', 'BELTRAMI']}) Plot random posterior predictive sub-samples. .. plot:: :context: close-figs >>> az.plot_ppc(data, num_pp_samples=30, random_seed=7) """ if group not in ("posterior", "prior"): raise TypeError("`group` argument must be either `posterior` or `prior`") for groups in (f"{group}_predictive", "observed_data"): if not hasattr(data, groups): raise TypeError(f'`data` argument must have the group "{groups}" for ppcplot') if kind.lower() not in ("kde", "cumulative", "scatter"): raise TypeError("`kind` argument must be either `kde`, `cumulative`, or `scatter`") if colors is None: colors = ["C0", "k", "C1"] if isinstance(colors, str): raise TypeError("colors should be a list with 3 items.") if len(colors) != 3: raise ValueError("colors should be a list with 3 items.") if color is not None: warnings.warn("color has been deprecated in favor of colors", FutureWarning) colors[0] = color if data_pairs is None: data_pairs = {} if backend is None: backend = rcParams["plot.backend"] backend = backend.lower() if backend == "bokeh" and animated: raise TypeError("Animation option is only supported with matplotlib backend.") observed_data = data.observed_data if group == "posterior": predictive_dataset = data.posterior_predictive if observed is None: observed = True elif group == "prior": predictive_dataset = data.prior_predictive if observed is None: observed = False if var_names is None: var_names = list(observed_data.data_vars) var_names = _var_names(var_names, observed_data, filter_vars) pp_var_names = [data_pairs.get(var, var) for var in var_names] pp_var_names = _var_names(pp_var_names, predictive_dataset, filter_vars) if flatten_pp is None: if flatten is None: flatten_pp = list(predictive_dataset.dims) else: flatten_pp = flatten if flatten is None: flatten = list(observed_data.dims) if coords is None: coords = {} else: coords = coords.copy() if labeller is None: labeller = BaseLabeller() if random_seed is not None: np.random.seed(random_seed) total_pp_samples = predictive_dataset.sizes["chain"] * predictive_dataset.sizes["draw"] if num_pp_samples is None: if kind == "scatter" and not animated: num_pp_samples = min(5, total_pp_samples) else: num_pp_samples = total_pp_samples if ( not isinstance(num_pp_samples, Integral) or num_pp_samples < 1 or num_pp_samples > total_pp_samples ): raise TypeError(f"`num_pp_samples` must be an integer between 1 and {total_pp_samples}.") pp_sample_ix = np.random.choice(total_pp_samples, size=num_pp_samples, replace=False) for key in coords.keys(): coords[key] = np.where(np.in1d(observed_data[key], coords[key]))[0] obs_plotters = filter_plotters_list( list( xarray_var_iter( observed_data.isel(coords), skip_dims=set(flatten), var_names=var_names, combined=True, dim_order=["chain", "draw"], ) ), "plot_ppc", ) length_plotters = len(obs_plotters) pp_plotters = [ tup for _, tup in zip( range(length_plotters), xarray_var_iter( predictive_dataset.isel(coords), var_names=pp_var_names, skip_dims=set(flatten_pp), combined=True, dim_order=["chain", "draw"], ), ) ] rows, cols = default_grid(length_plotters, grid=grid) ppcplot_kwargs = dict( ax=ax, length_plotters=length_plotters, rows=rows, cols=cols, figsize=figsize, animated=animated, obs_plotters=obs_plotters, pp_plotters=pp_plotters, predictive_dataset=predictive_dataset, pp_sample_ix=pp_sample_ix, kind=kind, alpha=alpha, colors=colors, jitter=jitter, textsize=textsize, mean=mean, observed=observed, observed_rug=observed_rug, total_pp_samples=total_pp_samples, legend=legend, labeller=labeller, group=group, animation_kwargs=animation_kwargs, num_pp_samples=num_pp_samples, backend_kwargs=backend_kwargs, show=show, ) # TODO: Add backend kwargs plot = get_plotting_function("plot_ppc", "ppcplot", backend) axes = plot(**ppcplot_kwargs) return axes