doubt.datasets.tehran_housing

  1"""Tehran housing data set.
  2
  3This data set is from the UCI data set archive, with the description being the original
  4description verbatim. Some feature names may have been altered, based on the
  5description.
  6"""
  7
  8import io
  9
 10import pandas as pd
 11
 12from .dataset import BASE_DATASET_DESCRIPTION, BaseDataset
 13
 14
 15class TehranHousing(BaseDataset):
 16    __doc__ = f"""
 17    Data set includes construction cost, sale prices, project variables, and economic
 18    variables corresponding to real estate single-family residential apartments in
 19    Tehran, Iran.
 20
 21    {BASE_DATASET_DESCRIPTION}
 22
 23    Features:
 24        start_year (int):
 25            Start year in the Persian calendar
 26        start_quarter (int)
 27            Start quarter in the Persian calendar
 28        completion_year (int)
 29            Completion year in the Persian calendar
 30        completion_quarter (int)
 31            Completion quarter in the Persian calendar
 32        V-1..V-8 (floats):
 33            Project physical and financial variables
 34        V-11-1..29-1 (floats):
 35            Economic variables and indices in time, lag 1
 36        V-11-2..29-2 (floats):
 37            Economic variables and indices in time, lag 2
 38        V-11-3..29-3 (floats):
 39            Economic variables and indices in time, lag 3
 40        V-11-4..29-4 (floats):
 41            Economic variables and indices in time, lag 4
 42        V-11-5..29-5 (floats):
 43            Economic variables and indices in time, lag 5
 44
 45    Targets:
 46        construction_cost (float)
 47        sale_price (float)
 48
 49    Source:
 50        https://archive.ics.uci.edu/ml/datasets/Residential+Building+Data+Set
 51
 52    Examples:
 53        Load in the data set::
 54
 55            >>> dataset = TehranHousing()
 56            >>> dataset.shape
 57            (371, 109)
 58
 59        Split the data set into features and targets, as NumPy arrays::
 60
 61            >>> X, y = dataset.split()
 62            >>> X.shape, y.shape
 63            ((371, 107), (371, 2))
 64
 65        Perform a train/test split, also outputting NumPy arrays::
 66
 67            >>> train_test_split = dataset.split(test_size=0.2, random_seed=42)
 68            >>> X_train, X_test, y_train, y_test = train_test_split
 69            >>> X_train.shape, y_train.shape, X_test.shape, y_test.shape
 70            ((288, 107), (288, 2), (83, 107), (83, 2))
 71
 72        Output the underlying Pandas DataFrame::
 73
 74            >>> df = dataset.to_pandas()
 75            >>> type(df)
 76            <class 'pandas.core.frame.DataFrame'>
 77    """
 78
 79    _url = (
 80        "https://archive.ics.uci.edu/ml/machine-learning-databases/"
 81        "00437/Residential-Building-Data-Set.xlsx"
 82    )
 83
 84    _features = range(107)
 85    _targets = [107, 108]
 86
 87    def _prep_data(self, data: bytes) -> pd.DataFrame:
 88        """Prepare the data set.
 89
 90        Args:
 91            data (bytes): The raw data
 92
 93        Returns:
 94            Pandas dataframe: The prepared data
 95        """
 96        # Convert the bytes into a file-like object
 97        xlsx_file = io.BytesIO(data)
 98
 99        # Load it into dataframe
100        cols = (
101            ["start_year", "start_quarter", "completion_year", "completion_quarter"]
102            + [f"V-{i}" for i in range(1, 9)]
103            + [f"V-{i}-{j}" for j in range(1, 6) for i in range(11, 30)]
104            + ["construction_cost", "sale_price"]
105        )
106        df = pd.read_excel(xlsx_file, skiprows=[0, 1], names=cols)
107        return df