nelmsal / landuse

a copy of a national land use regulatory index

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

landuse

a national land use regulatory index

This repository reproduces a workflow to predict land use with a combination of satellite imagery, administrative data, street topology. The imagery used in the model comes from Google Earth Engine, and this script produces the files.

  • help.R constains a function to create spatially lagged variables
  • extract.R allows us to convert raster imagery into tabular data
  • osmnx.ipynb constructs variables from the street network
  • left.R blends it all together to produce the left side of the regression
  • lag.R runs a lasso for feature selection and random forest for prediction

About

a copy of a national land use regulatory index


Languages

Language:Jupyter Notebook 98.6%Language:R 1.4%