tobler.model.glm

tobler.model.glm(source_df=None, target_df=None, raster='nlcd_2011', raster_codes=None, variable=None, formula=None, likelihood='poisson', force_crs_match=True, return_model=False)

Train a generalized linear model to predict polygon attributes based on the collection of pixel values they contain.

Parameters:

source_dfgeopandas.GeoDataFrame, required

geodataframe containing source original data to be represented by another geometry

target_dfgeopandas.GeoDataFrame, required

geodataframe containing target boundaries that will be used to represent the source data

rasterstr, required (default=”nlcd_2011”)

path to raster file that will be used to input data to the regression model. i.e. a coefficients refer to the relationship between pixel counts and population counts. Defaults to 2011 NLCD

raster_codeslist, required (default =[21, 22, 23, 24, 41, 42, 52])

list of integers that represent different types of raster cells. If no formula is given, the model will be fit from a linear combination of the logged count of each cell type listed here. Defaults to [21, 22, 23, 24, 41, 42, 52] which are informative land type cells from the NLCD

variablestr, required

name of the variable (column) to be modeled from the source_df

formulastr, optional

patsy-style model formula that specifies the model. Raster codes should be prefixed with “Type_”, e.g. “n_total_pop ~ -1 + np.log1p(Type_21) + np.log1p(Type_22)

likelihoodstr, {‘poisson’, ‘gaussian’, ‘neg_binomial’} (default = “poisson”)

the likelihood function used in the model

force_crs_matchbool

whether to coerce geodataframe and raster to the same CRS

return modelbool

whether to return the fitted model in addition to the interpolated geodataframe. If true, this will return (geodataframe, model)

Returns:

interpolatedgeopandas.GeoDataFrame

a new geopandas dataframe with boundaries from target_df and modeled attribute data from the source_df. If return_model is true, the function will also return the fitted regression model for further diagnostics