Thinking-data-science-a-data-science-practitioners-guide

26 2 Dimensionality Reduction
You may also do the correlation check through a program code like this:
feature_cols = df.columns[1:-1]
corr_values = df[feature_cols].corr()
indexes = np.tril_indices_from(corr_values)
for coord in zip(*indexes):
corr_values.iloc[coord[0], coord[1]] = np.NaN
corr_values = (corr_values
.stack()
.to_frame()
.reset_index()
.rename(columns={'level_0':'feature1',
'level_1':'feature2',
0:'correlation'}))
corr_values['abs_correlation'] =
corr_values.correlation.abs()
corr_values
The partial output on our dataset is shown in Fig. 2.4.
Fig. 2.4 Correlations in our dataset