1 2import numpy as np 3import matplotlib.pyplot as plt 4 5import statsmodels.api as sm 6 7 8# Necessary to make horizontal axis labels fit 9plt.rcParams['figure.subplot.bottom'] = 0.23 10 11data = sm.datasets.anes96.load_pandas() 12party_ID = np.arange(7) 13labels = ["Strong Democrat", "Weak Democrat", "Independent-Democrat", 14 "Independent-Independent", "Independent-Republican", 15 "Weak Republican", "Strong Republican"] 16 17# Group age by party ID. 18age = [data.exog['age'][data.endog == id] for id in party_ID] 19 20 21# Create a violin plot. 22fig = plt.figure() 23ax = fig.add_subplot(111) 24 25sm.graphics.violinplot(age, ax=ax, labels=labels, 26 plot_opts={'cutoff_val':5, 'cutoff_type':'abs', 27 'label_fontsize':'small', 28 'label_rotation':30}) 29 30ax.set_xlabel("Party identification of respondent.") 31ax.set_ylabel("Age") 32ax.set_title("US national election '96 - Age & Party Identification") 33 34 35# Create a bean plot. 36fig2 = plt.figure() 37ax = fig2.add_subplot(111) 38 39sm.graphics.beanplot(age, ax=ax, labels=labels, 40 plot_opts={'cutoff_val':5, 'cutoff_type':'abs', 41 'label_fontsize':'small', 42 'label_rotation':30}) 43 44ax.set_xlabel("Party identification of respondent.") 45ax.set_ylabel("Age") 46ax.set_title("US national election '96 - Age & Party Identification") 47 48 49# Create a jitter plot. 50fig3 = plt.figure() 51ax = fig3.add_subplot(111) 52 53plot_opts={'cutoff_val':5, 'cutoff_type':'abs', 'label_fontsize':'small', 54 'label_rotation':30, 'violin_fc':(0.8, 0.8, 0.8), 55 'jitter_marker':'.', 'jitter_marker_size':3, 'bean_color':'#FF6F00', 56 'bean_mean_color':'#009D91'} 57sm.graphics.beanplot(age, ax=ax, labels=labels, jitter=True, 58 plot_opts=plot_opts) 59 60ax.set_xlabel("Party identification of respondent.") 61ax.set_ylabel("Age") 62ax.set_title("US national election '96 - Age & Party Identification") 63 64 65# Create an asymmetrical jitter plot. 66ix = data.exog['income'] < 16 # incomes < $30k 67age = data.exog['age'][ix] 68endog = data.endog[ix] 69age_lower_income = [age[endog == id] for id in party_ID] 70 71ix = data.exog['income'] >= 20 # incomes > $50k 72age = data.exog['age'][ix] 73endog = data.endog[ix] 74age_higher_income = [age[endog == id] for id in party_ID] 75 76fig = plt.figure() 77ax = fig.add_subplot(111) 78 79plot_opts['violin_fc'] = (0.5, 0.5, 0.5) 80plot_opts['bean_show_mean'] = False 81plot_opts['bean_show_median'] = False 82plot_opts['bean_legend_text'] = r'Income < \$30k' 83plot_opts['cutoff_val'] = 10 84sm.graphics.beanplot(age_lower_income, ax=ax, labels=labels, side='left', 85 jitter=True, plot_opts=plot_opts) 86plot_opts['violin_fc'] = (0.7, 0.7, 0.7) 87plot_opts['bean_color'] = '#009D91' 88plot_opts['bean_legend_text'] = r'Income > \$50k' 89sm.graphics.beanplot(age_higher_income, ax=ax, labels=labels, side='right', 90 jitter=True, plot_opts=plot_opts) 91 92ax.set_xlabel("Party identification of respondent.") 93ax.set_ylabel("Age") 94ax.set_title("US national election '96 - Age & Party Identification") 95 96 97# Show all plots. 98plt.show() 99