Source code for evalml.demos.fraud

import os

from evalml.preprocessing import load_data


[docs]def load_fraud(n_rows=None, verbose=True, return_pandas=False): """Load credit card fraud dataset. The fraud dataset can be used for binary classification problems. Arguments: n_rows (int): Number of rows from the dataset to return verbose (bool): Whether to print information about features and labels Returns: Union[(ww.DataTable, ww.DataColumn), (pd.Dataframe, pd.Series)]: X and y """ currdir_path = os.path.dirname(os.path.abspath(__file__)) data_folder_path = os.path.join(currdir_path, "data") fraud_data_path = os.path.join(data_folder_path, "fraud_transactions.csv.gz") X, y = load_data(path=fraud_data_path, index="id", target="fraud", n_rows=n_rows, verbose=verbose) X = X.set_types({"provider": "Categorical", "region": "Categorical"}) if return_pandas: return X.to_dataframe(), y.to_series() return X, y