talkstats in r-bloggers

[bugman]

Anyone else see this? I thought I'd share since I found it pretty interesting...

http://www.r-bloggers.com/what-analytic-software-are-people-discussing/

 

[TheEcologist]

Nice find, good to see people are paying attention to us. What I think is nice is that this site is more all round, while the others seem heavily R biased. The other packages would hardly have showed up if it wasn't for TS.

 

[Dason]

I took a couple of minutes to recollect the data since I wasn't a fan of the stacked bar charts. If anybody else wants it here you go

dat <- structure(list(Site = c("SO", "TS", "CV", "SO", "TS", "CV", "SO", 
"TS", "CV", "SO", "TS", "CV"), Program = structure(c(1L, 1L, 
1L, 2L, 2L, 2L, 3L, 3L, 3L, 4L, 4L, 4L), .Label = c("R", "SAS", 
"SPSS", "Stata"), class = "factor"), value = c(23886L, 2012L, 
3234L, 980L, 826L, 187L, 186L, 1338L, 374L, 207L, 884L, 164L)), .Names = c("Site", 
"Program", "value"), row.names = c(NA, -12L), class = "data.frame")

Note that I thought it was more consistent to use the number of threads in each subforum for the program. It looks like the author of the article used the number of posts though - because if I try to recreate their plot:

qplot(Program, geom = "bar", fill =  Site, data = dat, weight = value)

... it's not quite the same

Edit: Here's the data using # of posts in the subforum for TS instead of # of threads

postdat <- structure(list(Site = c("SO", "TS", "CV", "SO", "TS", "CV", "SO", 
"TS", "CV", "SO", "TS", "CV"), Program = structure(c(1L, 1L, 
1L, 2L, 2L, 2L, 3L, 3L, 3L, 4L, 4L, 4L), .Label = c("R", "SAS", 
"SPSS", "Stata"), class = "factor"), value = c(23886, 10703, 
3234, 980, 2728, 187, 186, 3734, 374, 207, 3225, 164)), .Names = c("Site", 
"Program", "value"), row.names = c(NA, -12L), class = "data.frame")