Processing the sugar map

 Different implementations of the Sugarscpe agent-based model have different ways to get up the sugar map. SugarscapeCW used a built-in matrix (no space between each number) whereas the NetLogo model library uses an external text file with a blank space between each number. 

The easy way to translate between these two different format is to use the readr R package, something like this:

library(readr)

sugar_map1 <- read_fwf("~/tmp/sugar-map1.txt", fwf_widths(c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1), col_names = NULL))

write_delim(sugar_map1, "~/tmp/sugar-map-out.txt", delim = " ", col_names = FALSE)

Tidytext + stm

The new version of tidytext package now works with the stm package for topic modeling. This is a great new development.

Tidy Text Mining with R

Cool text mining book.

Another text analysis package

Quanteda seems to be a serious contender for analyzing textual data using R.