def fill_na_categorical_columns(self):
columns = [
'BsmtQual', 'BsmtCond', 'FireplaceQu', 'GarageQual', 'GarageCond',
'GarageFinish', 'GarageType', 'GarageYrBlt', 'MasVnrType',
'BsmtExposure', 'BsmtFinType1', 'BsmtFinType2', 'Fence'
]
for dataset in self.loop_datasets():
fill_nan_with_value(dataset, columns, 'NP')
def fill_with_zero(self):
columns = [
'BsmtFinSF1', 'BsmtFinSF2', 'BsmtUnfSF', 'TotalBsmtSF',
'GarageCars', 'GarageArea', 'BsmtFullBath', 'BsmtHalfBath',
'MasVnrArea'
]
for column in columns:
for dataset in self.loop_datasets():
fill_nan_with_value(dataset, [column], 0)
def test_fill_nan_with_value(self):
dataset = self.create_dummy_dataframe(10)
dataset.iloc[[1, 2, 3], dataset.columns.get_loc('a')] = np.NaN
fill_nan_with_value(dataset, ['a'], 0)
expected_dataset = self.create_dummy_dataframe(10)
expected_dataset.iloc[[1, 2, 3],
expected_dataset.columns.get_loc('a')] = 0
self.assertTrue(dataset.equals(expected_dataset))
dataset = self.create_dummy_dataframe(10)
dataset.iloc[[1, 2, 3], dataset.columns.get_loc('a')] = np.NaN
fill_nan_with_value(dataset, ['a'], 'TS')
expected_dataset = self.create_dummy_dataframe(10)
expected_dataset.iloc[[1, 2, 3],
expected_dataset.columns.get_loc('a')] = 'TS'
self.assertTrue(dataset.equals(expected_dataset))
def mode_fill(self, column):
train = self.data[0]
mode = train[column].mode()[0]
for dataset in self.loop_datasets():
fill_nan_with_value(dataset, [column], mode)
def fill_lot_frontage(self):
train = self.data[0]
mean = train['LotFrontage'].mean()
for dataset in self.loop_datasets():
fill_nan_with_value(dataset, ['LotFrontage'], mean)
def fill_mas_vnr_type(self):
for dataset in self.loop_datasets():
fill_nan_with_value(dataset, ['MasVnrType'], 'NP')