def PlotJointDist(pmf1, pmf2, thresh=0.8):
"""Plot the joint distribution of p_correct.
pmf1, pmf2: posterior distributions
thresh: lower bound of the range to be plotted
"""
def Clean(pmf):
"""Removes values below thresh."""
vals = [val for val in pmf.Values() if val < thresh]
[pmf.Remove(val) for val in vals]
Clean(pmf1)
Clean(pmf2)
pmf = thinkbayes2.MakeJoint(pmf1, pmf2)
thinkplot.Figure(figsize=(6, 6))
thinkplot.Contour(pmf, contour=False, pcolor=True)
thinkplot.Plot([thresh, 1.0], [thresh, 1.0],
color='gray',
alpha=0.2,
linewidth=4)
thinkplot.Save(root='sat_joint',
xlabel='p_correct Alice',
ylabel='p_correct Bob',
axis=[thresh, 1.0, thresh, 1.0],
formats=['pdf', 'eps'])
def PlotPosterior(suite, pcolor=False, contour=True):
"""Makes a contour plot.
suite: Suite that maps (mu, sigma) to probability
"""
thinkplot.Clf()
thinkplot.Contour(suite.GetDict(), pcolor=pcolor, contour=contour)
thinkplot.Save(root='variability_posterior_%s' % suite.name,
title='Posterior joint distribution',
xlabel='Mean height (cm)',
ylabel='Stddev (cm)')
def MakeContourPlot(suite):
"""Plots the posterior joint distribution as a contour plot.
suite: posterior joint distribution of location
"""
thinkplot.Contour(suite.GetDict(), contour=False, pcolor=True)
thinkplot.Save('paintball4',
xlabel='alpha',
ylabel='beta',
axis=[0, 30, 0, 20],
formats=FORMATS)
def PlotJointDist(self):
"""Makes a pcolor plot of the age-size joint distribution."""
thinkplot.Clf()
joint = self.cache.GetDistAgeSize()
thinkplot.Contour(joint, contour=False, pcolor=True)
thinkplot.Save(root='kidney8',
formats=FORMATS,
axis=[0, 41, -0.7, 1.31],
yticks=MakeLogTicks([0.2, 0.5, 1, 2, 5, 10, 20]),
xlabel='ages',
ylabel='diameter (cm, log scale)')
def MakeCrediblePlot(suite):
"""Makes a plot showing several two-dimensional credible intervals.
suite: Suite
"""
d = dict((pair, 0) for pair in suite.Values())
percentages = [75, 50, 25]
for p in percentages:
interval = suite.MaxLikeInterval(p)
for pair in interval:
d[pair] += 1
thinkplot.Contour(d, contour=False, pcolor=True)
pyplot.text(17, 4, '25', color='white')
pyplot.text(17, 15, '50', color='white')
pyplot.text(17, 30, '75')
thinkplot.Save('paintball5',
xlabel='alpha',
ylabel='beta',
formats=FORMATS)
# scatterplot of simulated records
for simulation in simulated_records:
plt.plot([item[0] for item in simulation], [item[1] for item in simulation], 'or')
plt.plot([pd.to_datetime(date) for date in dates], records, 'o')
plt.title("Simulated records")
plt.xlabel("year")
plt.ylabel("mph")
plt.show()
# pcolor plot of the simulated records
joint_estimate = thinkbayes2.Joint()
for simulation in simulated_records:
for date, record in simulation:
joint_estimate.Incr((date.value, round(record, 2)))
thinkplot.Contour(joint_estimate, contour=False, pcolor=True)
plt.plot([dates[0], pd.to_datetime('2041').value],[13.1094, 13.1094], 'r')
plt.plot([pd.to_datetime('2041').value, pd.to_datetime('2041').value], [12, 13.1094], 'r')
plt.plot(dates, records, 'bo')
plt.title("simulated records joint pmf")
thinkplot.show()
year2041 = find_closest(date_range, pd.to_datetime('2041'))
year2042 = find_closest(date_range, pd.to_datetime('2042'))
# calculate some summary statistics
pmf_2hour = joint_estimate.Conditional(0, 1, 13.11)
cdf_2hour = pmf_2hour.MakeCdf()
print("probability of 2-hour marathon by 2041: %f" % pmf_2hour.ProbLess(year2042.value))
print(pd.to_datetime(cdf_2hour.Value(.385445)))
print("year when we're 90% confident of having ran a 2-hour marathon " + str(pd.to_datetime(cdf_2hour.Value(.9))))
