Last year I published the above graphic, which then got converted into the below for the book London: The Information Capital. I have had many requests for the code I used to create the plot so here it is!
The data shown is the Office for National Statistics flow data. See here for the latest version. The file I used for the above can be downloaded here (it is >109 mb uncompressed so you need a decent computer to load/plot it all at once in R). You will also need this file of area (MSOA) codes and their co-ordinates. The code used is pasted below with comments above each segment. Good luck!
Load the flow data required – origin and destination points are needed.
input<-read.table("~/Dropbox/London_visualized_working/commute_flows/wu03ew_v1.csv", sep=",", header=T)
We only need the first 3 columns of the above
input<- input[,1:3] names(input)<- c("origin", "destination","total")
The UK Census file above didn’t have coordinates just area codes. Here is a lookup that provides those.
centroids<- read.csv("~/Dropbox/London_visualized_working/commute_flows/msoa_popweightedcentroids.csv") #Lots of joining to get the xy coordinates joined to the origin and then the destination points. or.xy<- merge(input, centroids, by.x="origin", by.y="Code") names(or.xy)<- c("origin", "destination", "trips", "o_name", "oX", "oY") dest.xy<- merge(or.xy, centroids, by.x="destination", by.y="Code") names(dest.xy)<- c("origin", "destination", "trips", "o_name", "oX", "oY","d_name", "dX", "dY")
Now for plotting with ggplot2.This first step removes the axes in the resulting plot.
xquiet<- scale_x_continuous("", breaks=NULL) yquiet<-scale_y_continuous("", breaks=NULL) quiet<-list(xquiet, yquiet)
Let’s build the plot. First we specify the dataframe we need, with a filter excluding flows of <10
ggplot(dest.xy[which(dest.xy$trips>10),], aes(oX, oY))+ #The next line tells ggplot that we wish to plot line segments. The "alpha=" is line transparency and used below geom_segment(aes(x=oX, y=oY,xend=dX, yend=dY, alpha=trips), col="white")+ #Here is the magic bit that sets line transparency - essential to make the plot readable scale_alpha_continuous(range = c(0.03, 0.3))+ #Set black background, ditch axes and fix aspect ratio theme(panel.background = element_rect(fill='black',colour='black'))+quiet+coord_equal()