Anscombe’s quartet
Part 1
For each of the four datasets…
- Compute the mean and variance of both x and y
- Compute the correlation coefficient between x and y
- Compute the linear regression line:
y=β0+β1x+ϵ
y
=
β
0
+
β
1
x
+
ϵ
(hint: use statsmodels and look at the Statsmodels notebook)
Part 2
Using Seaborn, visualize all four datasets.
hint: use sns.FacetGrid combined with plt.scatter
import random
import numpy as np
import scipy as sp
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import statsmodels.api as sm
import statsmodels.formula.api as smf
sns.set_context("talk")
anscombe = pd.read_csv('anscombe.csv')
print(anscombe.head())
# part 1
# 1.1
variables = 'xy'
print("\n1.1\nmean:")
for var in variables:
print(anscombe.groupby("dataset")[var].mean())
print("\nvariance:")
for var in variables:
print(anscombe.groupby("dataset")[var].var())
# 1.2
corrs = anscombe.groupby("dataset").corr()
print("\n1.2\ncorrelation coefficient between x and y")
for key in anscombe.groupby('dataset').indices:
print("%-6s" % key, corrs['x'][key]['y'])
# 1.3
print('\n1.3\nlinear regression:')
for key in anscombe.groupby('dataset').indices:
group = anscombe.groupby('dataset').get_group(key)
n = len(group)
is_train = np.random.rand(n)>-np.inf
train = group[is_train].reset_index(drop=True)
lin_model = smf.ols('y ~ x', train).fit()
print(key)
print(lin_model.summary() ,end='\n\n')
# part 2
ac = np.random.rand(len(anscombe)) < 0.5
print(smf.ols('y ~ x', anscombe[ac].reset_index(drop=True)).fit().summary())
sns.FacetGrid(anscombe, col="dataset").map(plt.scatter, 'x', 'y')
plt.show()