geopfa.extrapolation.load_2d_data

load_2d_data(gdf, *, value_col, z_value=None, z_tol=1e-06, test_size=0.2, seed=42)[source]

Prepare all arrays needed for 2D Gaussian process interpolation.

Parameters:

  • gdf (geopandas.GeoDataFrame) – Input GeoDataFrame containing coordinates and values.

  • value_col (str) – Column name containing the target variable.

  • z_value (float, optional) – Z slice to filter. If None, no Z filtering is applied.

  • z_tol (float, optional) – Tolerance for selecting points around z_value.

  • test_size (float, optional) – Fraction of non-missing data allocated to validation.

  • seed (int, optional) – Random seed for the train/validation split.

Returns:

dict – Dictionary with the same keys as the original implementation: includes standardized XY, grids, masks, pivoted arrays, and stats.