Extracting features from categorical variables

Many problems have explanatory variables that are categorical or nominal. A categorical variable can take one of a fixed set of values. For example, an application that predicts the salary for a job might use categorical variables such as the city in which the position is located. Categorical variables are commonly encoded using one-of-k encoding, or one-hotencoding, in which the explanatory variable is represented using one binary feature for each of its possible values.

For example, let's assume our model has a city variable that can take one of three values: New York, San Francisco, or Chapel Hill. One-hot encoding represents the variable using one binary feature for each of the three possible cities. scikit-learn's DictVectorizer class is a transformer that can be used to one-hot encode categorical features:

# In[1]:
from sklearn.feature_extraction import DictVectorizer
onehot_encoder = DictVectorizer()
X = [
    {'city': 'New York'},
   ...