Testing several scenarios with the new probplot functionality

proboplot_test.py

import matplotlib.pyplot as plt
import scipy.stats as stats
import statsmodels.api as sm

# longley dataset
data = sm.datasets.longley.load()
data.exog = sm.add_constant(data.exog)
model = sm.OLS(data.endog, data.exog)
mod_fit = model.fit()
res = mod_fit.resid


# T-dist
pp = sm.ProbPlot(res, dist=stats.t, distargs=(4,))
for line in ['r', 'q', '45', 's']:
    fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, ncols=1)
    pp.qqplot(ax1, line=line)
    pp.ppplot(ax2, line=line)
    pp.probplot(ax3, line=line)
    fig.suptitle('T+distargs with line='+line)
    plt.show()

pp = sm.ProbPlot(res, dist=stats.t, distargs=(4,), loc=3, scale=10)
for line in ['r', 'q', '45', 's']:
    fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, ncols=1)
    pp.qqplot(ax1, line=line)
    pp.ppplot(ax2, line=line)
    pp.probplot(ax3, line=line)
    fig.suptitle('T+distargs+log+scale with line='+line)
    plt.show()

pp = sm.ProbPlot(res, dist=stats.t, fit=True)
for line in ['r', 'q', '45', 's']:
    fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, ncols=1)
    pp.qqplot(ax1, line=line)
    pp.ppplot(ax2, line=line)
    pp.probplot(ax3, line=line)
    fig.suptitle('T+fit with line='+line)
    plt.show()


# Normal dist
pp = sm.ProbPlot(res, dist=stats.norm)
for line in ['r', 'q', '45', 's']:
    fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, ncols=1)
    pp.qqplot(ax1, line=line)
    pp.ppplot(ax2, line=line)
    pp.probplot(ax3, line=line)
    fig.suptitle('Norm with line='+line)
    plt.show()

pp = sm.ProbPlot(res, dist=stats.norm, loc=res.mean(), scale=res.std())
for line in ['r', 'q', '45', 's']:
    fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, ncols=1)
    pp.qqplot(ax1, line=line)
    pp.ppplot(ax2, line=line)
    pp.probplot(ax3, line=line)
    fig.suptitle('Norm+log+scale with line='+line)
    plt.show()

pp = sm.ProbPlot(res, dist=stats.norm, fit=True)
for line in ['r', 'q', '45', 's']:
    fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, ncols=1)
    pp.qqplot(ax1, line=line)
    pp.ppplot(ax2, line=line)
    pp.probplot(ax3, line=line)
    fig.suptitle('Norm+fit with line='+line)
    plt.show()