Etiquetas

, , ,

Some days ago James Cheshire published the post Mapping Flows in R. I have implemented an alternative (faster) version using data.table to read and join the datasets (and lattice to display the results). If you are new to data.table you should read this wiki and this cheatsheet.

This is the code:


### DATA SECTION
library(data.table)
## Read data with 'data.table::fread'
input <- fread("wu03ew_v1.csv", select = 1:3)
setnames(input, 1:3, new = c("origin", "destination","total"))
## Coordinates
centroids <- fread("msoa_popweightedcentroids.csv")
## 'Code' is the key to be used in the joins
setkey(centroids, Code)
## Key of centroids matches `origin` in `input`
origin <- centroids[input[,.(origin, total)]]
setnames(origin, c('East', 'North'), c('xstart', 'ystart'))
## Key of centroids matches `destination` in `input`
destination <- centroids[input[,.(destination)]]
setnames(destination, c('East', 'North'), c('xend', 'yend'))
## Bind both results
trajects <- cbind(origin, destination)
### GRAPHICS SECTION
library(lattice)
library(classInt)
## Background set to black
myTheme <- simpleTheme()
myTheme$background$col <- 'black'
## Palette and classes
nClasses <- 5
pal <- colorRampPalette(c('gray70', 'white'))(nClasses)
classes <- classIntervals(trajects[total > 10, total],
n = nClasses, style = 'quantile')
classes <- findCols(classes)
xyplot(North ~ East, data = centroids,
pch = '.', col = 'lightgray',
aspect = 'iso',
par.settings = myTheme,
panel = function(…){
## panel.xyplot displays the 'centroids'
panel.xyplot(…)
## panel.segments displays the lines using a `data.table`
## query.
trajects[total > 10,
panel.segments(xstart, ystart, xend, yend,
col = pal[classes],
alpha = 0.05, lwd = 0.3)
]
})

view raw

mappingFlows.R

hosted with ❤ by GitHub

This is the result:

london

Related articles