/spatialpy

formally skspatial but someone at scikit-learn took the name and it messed up people

Primary LanguagePython

Alt text

skspatial

Simple functions for geospatial interpolation using sklearn's KNN machine learning algorithm, simple scipy interpolation routines ie. "linear", or "cubic" and now pykrige for kriging functions.

Simply load a projected point shapefile with geopandas as a GeoDataFrame, and use skspatial to create interpolated rasters and countor shapefiles that you can bring into your favorite mapping application such as QGIS.

Currently in development by:

Alt text

Written by Ross Kushnereit, Intera Geoscience & Engineering Solutions: Austin, TX https://www.intera.com/

Installation

skspatial supports Python 3.6

    $ git clone https://github.com/rosskush/skspatial.git
    $ cd skspatial
    $ python setup.py install

Reqiuerments

skspatial in its current state reqiures at a minimum

geopandas

rasterio

and sklearn

Refrences

http://chris35wills.github.io/gridding_data/

https://timogrossenbacher.ch/2018/03/categorical-spatial-interpolation-with-r/

http://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsRegressor.html#sklearn.neighbors.KNeighborsRegressor