This kind of quantitative transformation is used to create quantiles. In this case, the quantitative feature values will be the transformed ordered variable. This approach is not a good choice for linear regression, but it might work well for learning algorithms that respond effectively when using ordered/categorical variables.

The following code applies this kind of transformation to the Fare feature:

# Binarizing the features by binning them into quantiles
df_titanic_data['Fare_bin'] = pd.qcut(df_titanic_data['Fare'], 4)

if keep_binary:
    df_titanic_data = pd.concat(
        [df_titanic_data, pd.get_dummies(df_titanic_data['Fare_bin']).rename(columns=lambda x: 'Fare_' + str(x))],
        axis=1)