Skip to content

utilities

predspot.utilities

Utilities Module

Helpers used across Predspot: a PandasFeatureUnion that keeps DataFrames (and their index) when combining feature transformers, and a GeoJSON contour export for density maps.

PandasFeatureUnion 

PandasFeatureUnion(transformer_list)

Bases: TransformerMixin, BaseEstimator

Concatenate the DataFrame outputs of several transformers column-wise.

Unlike sklearn.pipeline.FeatureUnion, the transformers' outputs are aligned on their index and returned as a DataFrame. Rows with missing values after alignment (e.g. warm-up rows of lag features) are dropped.

Parameters:

Name Type Description Default
transformer_list list

(name, transformer) pairs.

 required
Source code in src/predspot/utilities.py 
34
35
def __init__(self, transformer_list):
    self.transformer_list = transformer_list

merge_dataframes_by_column staticmethod 

merge_dataframes_by_column(outputs)

Align a list of DataFrames on their index and concatenate columns.

Parameters:

Name Type Description Default
outputs list

DataFrames returned by the transformers.

 required

Returns:

Type Description
DataFrame

The merged features without missing rows.
Source code in src/predspot/utilities.py 
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
@staticmethod
def merge_dataframes_by_column(outputs):
    """
    Align a list of DataFrames on their index and concatenate columns.

    Args:
        outputs (list): DataFrames returned by the transformers.

    Returns:
        pandas.DataFrame: The merged features without missing rows.
    """
    if not outputs:
        raise ValueError("PandasFeatureUnion has no transformers.")
    logger.debug("Merging %d feature blocks", len(outputs))
    return pd.concat(outputs, axis="columns").dropna()

contour_geojson 

contour_geojson(y, bbox, resolution, cmin, cmax)

Export a density surface as filled GeoJSON contours.

Requires the optional dependency geojsoncontour (pip install predspot[contour]).

Parameters:

Name Type Description Default
y Series

Values indexed by the positional index of the full point grid returned by create_gridpoints (before clipping), i.e. the places index.

 required
bbox GeoDataFrame

Study area used to build the grid.

 required
resolution float

Grid resolution in kilometers (same as the grid).

 required
cmin float

Lowest contour level.

 required
cmax float

Highest contour level.

 required

Returns:

Type Description
str

GeoJSON string with the contour polygons.
Source code in src/predspot/utilities.py 
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
def contour_geojson(y, bbox, resolution, cmin, cmax):
    """
    Export a density surface as filled GeoJSON contours.

    Requires the optional dependency ``geojsoncontour``
    (``pip install predspot[contour]``).

    Args:
        y (pandas.Series): Values indexed by the positional index of the
            full point grid returned by
            [`create_gridpoints`][predspot.crime_mapping.create_gridpoints] (before
            clipping), i.e. the ``places`` index.
        bbox (GeoDataFrame): Study area used to build the grid.
        resolution (float): Grid resolution in kilometers (same as the grid).
        cmin (float): Lowest contour level.
        cmax (float): Highest contour level.

    Returns:
        str: GeoJSON string with the contour polygons.
    """
    try:
        import geojsoncontour
    except ImportError as exc:  # pragma: no cover - optional dependency
        raise ImportError(
            "contour_geojson requires the optional dependency "
            "`geojsoncontour`: pip install predspot[contour]"
        ) from exc
    import matplotlib

    matplotlib.use("Agg")
    import matplotlib.pyplot as plt

    from predspot.crime_mapping import KM_PER_DEG_LAT, KM_PER_DEG_LON, _check_bbox, _wgs84_bounds

    _check_bbox(bbox)
    b_w, b_s, b_e, b_n = _wgs84_bounds(bbox)
    nlon = max(int(np.ceil((b_e - b_w) / (resolution / KM_PER_DEG_LON))), 2)
    nlat = max(int(np.ceil((b_n - b_s) / (resolution / KM_PER_DEG_LAT))), 2)
    lonv, latv = np.meshgrid(np.linspace(b_w, b_e, nlon), np.linspace(b_s, b_n, nlat))
    Z = np.full(lonv.size, -999.0)
    Z[np.asarray(y.index, dtype=int)] = y.values
    Z = Z.reshape(lonv.shape)

    fig, axes = plt.subplots()
    contourf = axes.contourf(lonv, latv, Z, levels=np.linspace(cmin, cmax, 25), cmap="Spectral_r")
    geojson = geojsoncontour.contourf_to_geojson(contourf=contourf, fill_opacity=0.5)
    plt.close(fig)
    return geojson