Make a violin plot of each dataset in the data sequence.
A violin plot is a boxplot combined with a kernel density estimate of the probability density function per point.
Parameters: | data : sequence of ndarrays
ax : Matplotlib AxesSubplot instance, optional
labels : list of str, optional
positions : array_like, optional
side : {‘both’, ‘left’, ‘right’}, optional
show_boxplot : bool, optional
plot_opts : dict, optional
|
---|---|
Returns: | fig : Matplotlib figure instance
|
See also
Notes
The appearance of violins can be customized with plot_opts. If customization of boxplot elements is required, set show_boxplot to False and plot it on top of the violins by calling the Matplotlib boxplot function directly. For example:
violinplot(data, ax=ax, show_boxplot=False)
ax.boxplot(data, sym='cv', whis=2.5)
It can happen that the axis labels or tick labels fall outside the plot area, especially with rotated labels on the horizontal axis. With Matplotlib 1.1 or higher, this can easily be fixed by calling ax.tight_layout(). With older Matplotlib one has to use plt.rc or plt.rcParams to fix this, for example:
plt.rc('figure.subplot', bottom=0.25)
violinplot(data, ax=ax)
References
J.L. Hintze and R.D. Nelson, “Violin Plots: A Box Plot-Density Trace Synergism”, The American Statistician, Vol. 52, pp.181-84, 1998.
Examples
We use the American National Election Survey 1996 dataset, which has Party Identification of respondents as independent variable and (among other data) age as dependent variable.
>>> data = sm.datasets.anes96.load_pandas()
>>> party_ID = np.arange(7)
>>> labels = ["Strong Democrat", "Weak Democrat", "Independent-Democrat",
... "Independent-Indpendent", "Independent-Republican",
... "Weak Republican", "Strong Republican"]
Group age by party ID, and create a violin plot with it:
>>> plt.rcParams['figure.subplot.bottom'] = 0.23 # keep labels visible
>>> age = [data.exog['age'][data.endog == id] for id in party_ID]
>>> fig = plt.figure()
>>> ax = fig.add_subplot(111)
>>> sm.graphics.violinplot(age, ax=ax, labels=labels,
... plot_opts={'cutoff_val':5, 'cutoff_type':'abs',
... 'label_fontsize':'small',
... 'label_rotation':30})
>>> ax.set_xlabel("Party identification of respondent.")
>>> ax.set_ylabel("Age")
>>> plt.show()
(Source code, png, hires.png, pdf)