# Convert our categorical values to numeric feature vectors
feature_vecs = numpy.array([
        votes[c].cat.codes 
        for c in votes.columns 
        if c != label_col]).T
print(feature_vecs.shape)