def run_titanic_analysis(filename):
"""Data pipeline and predictions.
Parameters
----------
filename: str
Path to the Titanic CSV input data
"""
logging.info('Starting the data analysis pipeline')
processed_data = (
pd.read_csv(filename, usecols=['Name', 'Sex', 'Age', 'Survived'])
.pipe(lambda df: df.fillna({'Age': df.Age.median(), }))
.pipe(lambda df: df.astype({'Age': 'float64',
'Name': 'object',
'Sex': 'category',
'Survived': 'int64'}))
.pipe(data.extract_title)
)
X_train, X_test, y_train, y_test = models.data_preparation(processed_data,
test_size=0.2,
random_state=0)
models.run_majority_vote(X_train, X_test, y_train, y_test)
models.run_logistic_regression(X_train, X_test, y_train, y_test)
logging.info('The data analysis pipeline has terminated')
return
def test_run_logistic_regression():
processed_data = (pd.read_csv(
validation_data,
usecols=['Name', 'Sex', 'Age',
'Survived']).pipe(lambda df: df.fillna({
'Age': df.Age.median(),
})).pipe(lambda df: df.astype({
'Age': 'float64',
'Name': 'object',
'Sex': 'category',
'Survived': 'int64'
})).pipe(data.extract_title))
X_train, X_test, y_train, y_test = models.data_preparation(processed_data,
test_size=0.2,
random_state=0)
majority_vote = models.run_majority_vote(X_train, X_test, y_train, y_test)
linear_regression = models.run_logistic_regression(X_train, X_test,
y_train, y_test)
accuracy_majority_vote = accuracy_score(
y_true=y_test, y_pred=majority_vote.predict(X_test))
accuracy_linear_regression = accuracy_score(
y_true=y_test, y_pred=linear_regression.predict(X_test))
assert accuracy_linear_regression > accuracy_majority_vote