step_4_1.py

cur_data = price_data.copy()
cur_data = cur_data.merge(rental_data, on="id")
cur_data = cur_data.merge(location_data, on="id")

X_train, X_test, y_train, y_test = \
    custom_train_test_split(cur_data)

pass_cols = ["is_brooklyn", "density"]
drop_cols = ["year", "geometry", "zipcode"]
one_hot_cols = ["month"]
poly_cols = rental_data.columns.drop("id").tolist()

mentioned_cols = [ 
    "id", "price", 
    *one_hot_cols, *drop_cols, 
    *poly_cols, *pass_cols
]

assert sorted(cur_data.columns) == sorted(mentioned_cols)

cur_one_hot = OneHotEncoder(categories="auto")

cur_poly_feats = PolynomialFeatures(
    degree=2, include_bias=False, interaction_only=True
)

cur_sub_pipeline = Pipeline([
    ("reciprocal", ReciprocalFeatures()),
    ("polynomial", cur_poly_feats),
    ("cancel", VarianceThreshold()),
    ("clean", CleanFeatures()),
    ("box_cox", PowerTransformer(method="box-cox"))
])

cur_col_transformers = [
    ("one_hot", cur_one_hot, one_hot_cols),
    ("poly_feats", cur_sub_pipeline, poly_cols),
    ("passthrough", PassThroughTransformer(), pass_cols)
]

cur_transformer = ColumnTransformer(cur_col_transformers)

cur_selector = SelectFromModel(
    LogTransformedTargetRegressor(
        LassoCV(cv=4, n_jobs=-1, max_iter=5e4)
    ), threshold=5e-4
)

cur_regressor = LogTransformedTargetRegressor(
    ElasticNetCV(cv=4, n_jobs=-1, max_iter=5e4)
)

cur_pipeline = Pipeline([
    ("transformer", cur_transformer),
    ("scalar", StandardScaler()),
    ("selector", cur_selector),
    ("regressor", cur_regressor)
])

cur_pipeline.fit(X_train, y_train)
cur_pipeline.score(X_test, y_test)

This uses the following code:

• https://gist.github.com/djsegal/237d96c0eef26002c4aaf6c6ef957374

And produces the following csv of feature importances:

• https://gist.github.com/djsegal/49ea120e6754556ceba4aa8a6087bc80