def MakeHists(live):
"""Plot Hists for live births
live: DataFrame
others: DataFrame
"""
hist = mystats.Hist(live.birthwgt_lb, label="birthwgt_lb")
myplots.Hist(hist)
myplots.Save(root="first_wgt_lb_hist",
xlabel="pounds",
ylabel="frequency",
axis=[-1, 14, 0, 3200])
hist = mystats.Hist(live.birthwgt_oz, label="birthwgt_oz")
myplots.Hist(hist)
myplots.Save(root="first_wgt_oz_hist",
xlabel="ounces",
ylabel="frequency",
axis=[-1, 16, 0, 1200])
hist = mystats.Hist(np.floor(live.agepreg), label="agepreg")
myplots.Hist(hist)
myplots.Save(root="first_agepreg_hist", xlabel="years", ylabel="frequency")
hist = mystats.Hist(live.prglngth, label="prglngth")
myplots.Hist(hist)
myplots.Save(root="first_prglngth_hist",
xlabel="weeks",
ylabel="frequency",
axis=[-1, 53, 0, 5000])
def RunModel(self):
n = sum(self.data)
values = [1, 2, 3, 4, 5, 6]
rolls = np.random.choice(values, n, replace=True)
hist = mystats.Hist(rolls)
freqs = hist.Freqs(values)
return freqs
def RunModel(self):
heads, tails = self.data
n = heads + tails
sample = [random.choice("HT") for _ in range(n)]
hist = mystats.Hist(sample)
data = hist["H"], hist["T"]
return data
def MakeComparison(firsts, others):
"""Plots histograms of pregnancy length for first babies and others.
firsts: DataFrame
others: DataFrame
"""
first_hist = mystats.Hist(firsts.prglngth, label="first")
other_hist = mystats.Hist(others.prglngth, label="other")
width = 0.45
myplots.PrePlot(2)
myplots.Hist(first_hist, align="right", width=width)
myplots.Hist(other_hist, align="left", width=width)
myplots.Save(
root="first_nsfg_hist",
title="Histogram",
xlabel="weeks",
ylabel="frequency",
axis=[27, 46, 0, 2700],
)
def PrintExtremes(live):
"""Plots the histogram of pregnancy lengths and prints the extremes.
live: DataFrame of live births
"""
hist = mystats.Hist(live.prglngth)
myplots.Hist(hist, label="live births")
myplots.Save(root="first_nsfg_hist_live",
title="Histogram",
xlabel="weeks",
ylabel="frequency")
print("Shortest lengths:")
for weeks, freq in hist.Smallest(10):
print(weeks, freq)
print("Longest lengths:")
for weeks, freq in hist.Largest(10):
print(weeks, freq)
def ChiSquared(self, lengths):
hist = mystats.Hist(lengths)
observed = np.array(hist.Freqs(self.values))
expected = self.expected_probs * len(lengths)
stat = sum((observed - expected)**2 / expected)
return stat