Exporting matrix from ArcGIS (so readable in R)

Question

I would like to eventually create a shapefile in ArcGIS and then extract it as a matrix to be read by a spatial package I use in R. By matrix I actually mean a raster file.

My first step is to create a simple binary matrix of a map (UK for example) with 1 - water and 2- land. As a newcomer to both coding and GIS, I am struggling to come up with a solution. - converting to raster and exporting as text file was exactly what I needed.

I have been looking into spatial weights matrices - does this sound like the right way forward?

This will eventually need to be executed multiple times with different layers and different data sets - so will need a Python script for it. Right now I am just happy with exporting one text file for one layer.

Current problem: Although I can change the cell size - I really need to be able to alter the number of columns and rows. I want it to be a 300 x 300 matrix to save any complicated maths later on.

Background: The package I use (not yet on CRAN) looks at spatial points and displays them on Google Maps with predicted 'sources'. I would like to develop this and reducing the search area (give more informed priors). For example 'don't search in water'. To do this I need a nice matrix from ArcGIS.

Any ideas?

Answer 1

Thanks everyone.

Converting the layer to a raster and exporting as a text file was exactly what I needed for the time being. I used the convert to Raster function in the toolbox. .

You can then use the raster to ASCII function to save the data as a txt file which can then be read by my r package.

Answer 2

You are over complicating this a bit. In ArcGIS you can save your raster(s) as tiff or img formats. In R you can use the raster package to read the rasters (using raster(), stack() or brick() functions) and covert to a matrix or array in a single step. If you have a single raster then you would use the raster() function. However, if you are working with a stack (multi band) raster object then you would use the stack() or brick() functions and coerce using as.array().

library(raster)

#Create an example raster, would result in same object as reading raster from disk
r <- raster(matrix(rnorm(25),5))

# Coerce raster to matrix
m <- as.matrix(r)

# In a single step
( m <- as.matrix(raster(matrix(rnorm(25),5))) )

# Here is an array example
r <- stack(raster(matrix(rnorm(25),5)), raster(matrix(rnorm(25),5)))
( a <- as.array(r) )