arviz_stats.qds

Contents

arviz_stats.qds#

arviz_stats.qds(data, dim=None, group='posterior', var_names=None, filter_vars=None, coords=None, nquantiles=100, binwidth=None, dotsize=1, stackratio=1, **kwargs)[source]#

Compute the marginal quantile dots.

For details see [1] and check the EABM chapter on Visualization of Random Variables with ArviZ.

Parameters:
dataarray_like, xarray.DataArray, xarray.Dataset, xarray.DataTree, DataArrayGroupBy, DatasetGroupBy, or idata-like

Input data. It will have different pre-processing applied to it depending on its type:

  • array-like: call array layer within arviz-stats.

  • xarray object: apply dimension aware function to all relevant subsets

  • others: passed to arviz_base.convert_to_dataset then treated as xarray.Dataset. This option is discouraged due to needing this conversion which is completely automated and will be needed again in future executions or similar functions.

    It is recommended to first perform the conversion manually and then call arviz_stats.kde. This allows controlling the conversion step and inspecting its results.

dimsequence of hashable, optional

Dimensions to be reduced when computing the quantile dots Default rcParams["data.sample_dims"].

grouphashable, default “posterior”

Group on which to compute the quantile dots

var_namesstr or list of str, optional

Names of the variables for which the quantile dots should be computed.

filter_vars{None, “like”, “regex”}, default None
coordsdict, optional

Dictionary of dimension/index names to coordinate values defining a subset of the data for which to perform the computation.

binwidthfloat, optional

Width of the bin for the dots.

dotsizefloat, default 1

The size of the dots relative to the bin width. The default makes dots be just about as wide as the bin width.

stackratiofloat, default 1

The distance between the center of the dots in the same stack relative to the bin height. The default makes dots in the same stack just touch each other.

**kwargsany, optional

Forwarded to the array or dataarray interface for quantile dots.

Returns:
ndarray, xarray.DataArray, xarray.Dataset, xarray.DataTree

Requested QDs of the provided input. The xarray objects will have a qds_dim dimension and a plot_axis dimension with coordinates “x”, and “y”.

See also

arviz_stats.ecdf, arviz_stats.histogram, arviz_stats.kde

Alternative visual summaries for marginal distributions

arviz_plots.plot_dist

References

[1]

Kay M, Kola T, Hullman JR, and Munson SA. When (ish) is My Bus?: User-centered Visualizations of Uncertainty in Everyday, Mobile Predictive Systems. In Proceedings of the 2016 CHI Conference Association for Computing Machinery. 2016. https://doi.org/10.1145/2858036.2858558

Examples

Calculate the QDs of a Normal random variable:

In [1]: import arviz_stats as azs
   ...: import numpy as np
   ...: data = np.random.default_rng().normal(size=2000)
   ...: azs.qds(data)
   ...: 
Out[1]: 
(array([-2.35654682, -1.98214005, -1.98214005, -1.62626319, -1.62626319,
        -1.62626319, -1.38161377, -1.38161377, -1.38161377, -1.38161377,
        -1.12575565, -1.12575565, -1.12575565, -1.12575565, -1.12575565,
        -0.88593784, -0.88593784, -0.88593784, -0.88593784, -0.88593784,
        -0.88593784, -0.65648948, -0.65648948, -0.65648948, -0.65648948,
        -0.65648948, -0.65648948, -0.65648948, -0.4343695 , -0.4343695 ,
        -0.4343695 , -0.4343695 , -0.4343695 , -0.4343695 , -0.4343695 ,
        -0.4343695 , -0.22670758, -0.22670758, -0.22670758, -0.22670758,
        -0.22670758, -0.22670758, -0.22670758, -0.22670758, -0.01546457,
        -0.01546457, -0.01546457, -0.01546457, -0.01546457, -0.01546457,
        -0.01546457, -0.01546457, -0.01546457,  0.19720502,  0.19720502,
         0.19720502,  0.19720502,  0.19720502,  0.19720502,  0.19720502,
         0.19720502,  0.40432871,  0.40432871,  0.40432871,  0.40432871,
         0.40432871,  0.40432871,  0.40432871,  0.40432871,  0.61181511,
         0.61181511,  0.61181511,  0.61181511,  0.61181511,  0.61181511,
         0.61181511,  0.82616068,  0.82616068,  0.82616068,  0.82616068,
         0.82616068,  0.82616068,  1.0478598 ,  1.0478598 ,  1.0478598 ,
         1.0478598 ,  1.0478598 ,  1.27581663,  1.27581663,  1.27581663,
         1.27581663,  1.5041899 ,  1.5041899 ,  1.5041899 ,  1.76303412,
         1.76303412,  1.76303412,  2.0909525 ,  2.0909525 ,  2.63373323]),
 array([0.09954169, 0.09954169, 0.29862506, 0.09954169, 0.29862506,
        0.49770843, 0.09954169, 0.29862506, 0.49770843, 0.6967918 ,
        0.09954169, 0.29862506, 0.49770843, 0.6967918 , 0.89587517,
        0.09954169, 0.29862506, 0.49770843, 0.6967918 , 0.89587517,
        1.09495854, 0.09954169, 0.29862506, 0.49770843, 0.6967918 ,
        0.89587517, 1.09495854, 1.29404191, 0.09954169, 0.29862506,
        0.49770843, 0.6967918 , 0.89587517, 1.09495854, 1.29404191,
        1.49312528, 0.09954169, 0.29862506, 0.49770843, 0.6967918 ,
        0.89587517, 1.09495854, 1.29404191, 1.49312528, 0.09954169,
        0.29862506, 0.49770843, 0.6967918 , 0.89587517, 1.09495854,
        1.29404191, 1.49312528, 1.69220865, 0.09954169, 0.29862506,
        0.49770843, 0.6967918 , 0.89587517, 1.09495854, 1.29404191,
        1.49312528, 0.09954169, 0.29862506, 0.49770843, 0.6967918 ,
        0.89587517, 1.09495854, 1.29404191, 1.49312528, 0.09954169,
        0.29862506, 0.49770843, 0.6967918 , 0.89587517, 1.09495854,
        1.29404191, 0.09954169, 0.29862506, 0.49770843, 0.6967918 ,
        0.89587517, 1.09495854, 0.09954169, 0.29862506, 0.49770843,
        0.6967918 , 0.89587517, 0.09954169, 0.29862506, 0.49770843,
        0.6967918 , 0.09954169, 0.29862506, 0.49770843, 0.09954169,
        0.29862506, 0.49770843, 0.09954169, 0.29862506, 0.09954169]),
 array(0.09954169))

Calculate the QDs for specific variables:

In [2]: import arviz_base as azb
   ...: dt = azb.load_arviz_data("centered_eight")
   ...: azs.qds(dt, var_names=["mu", "theta"])
   ...: 
Out[2]: 
<xarray.DataTree 'posterior'>
Group: /posterior
    Dimensions:       (plot_axis: 2, qd_dim: 100, school: 8)
    Coordinates:
      * plot_axis     (plot_axis) <U1 8B 'x' 'y'
        radius_mu     float64 8B 0.3682
      * school        (school) <U16 512B 'Choate' 'Deerfield' ... 'Mt. Hermon'
        radius_theta  (school) float64 64B 0.7168 0.5463 0.6846 ... 0.5679 0.7148
    Dimensions without coordinates: qd_dim
    Data variables:
        mu            (plot_axis, qd_dim) float64 2kB -5.17 -3.397 ... 0.3682 0.3682
        theta         (plot_axis, school, qd_dim) float64 13kB -7.762 ... 0.7148

Calculate the QDs also over the school dimension (for variables where present):

In [3]: azs.qds(dt, dim=["chain", "draw", "school"])
Out[3]: 
<xarray.DataTree 'posterior'>
Group: /posterior
    Dimensions:       (plot_axis: 2, qd_dim: 100)
    Coordinates:
      * plot_axis     (plot_axis) <U1 8B 'x' 'y'
        radius_mu     float64 8B 0.3682
        radius_theta  float64 8B 0.6464
        radius_tau    float64 8B 0.3271
    Dimensions without coordinates: qd_dim
    Data variables:
        mu            (plot_axis, qd_dim) float64 2kB -5.17 -3.397 ... 0.3682 0.3682
        theta         (plot_axis, qd_dim) float64 2kB -11.06 -7.544 ... 0.6464
        tau           (plot_axis, qd_dim) float64 2kB 1.199 1.199 ... 0.3271 0.3271
Contents