analyze_data.R

## top 20 packages
tail(sort(table(downloads$package)), 20)
## create a data table of packages and download numbers
package_downloads<-as.data.frame(table(downloads$package))
names(package_downloads)<- c("package","downloads")
library(dplyr)
# sort by download number
package_downloads <- arrange(package_downloads, desc(downloads))

# add rank column
barplot(head(package_downloads$downloads,10), names.arg=head(package_downloads$package,10))
package_downloads <- mutate(package_downloads, rank=c(1:9704))

# reverse dependencies and reverse imports
library(tools)
package_downloads <- mutate(package_downloads, rv_dp=lengths(package_dependencies(package_downloads$package, which=c("Depends", "Imports", "LinkingTo"), reverse=TRUE)))

save(package_downloads, file="package_downloads.Rdata")

##operating systems
os<-as.data.frame(table(downloads$r_os))

##r versions
r_vers<-as.data.frame(table(downloads$r_version))

substr(r_vers$Var1,1,1)
gsub(".","", r_vers$Var1)
r_vers$Var1


## countries
countries <- as.data.frame(table(downloads$country))
names(countries) <- c("country", "freq")
countries <- arrange(countries, desc(freq))

## countries and packages
library(dplyr)
popular <- data.frame(matrix(ncol=10, nrow=length(countries$country)))
for(i in 1:length(countries$country)){
  message(i)
  a <- downloads[country==countries$country[i],]
  b <- arrange(as.data.frame(table(a$package)), desc(Freq))
  c <- as.character(head(b[,1],10))
  popular[i,] <- c
}
popular <- cbind(countries$country, cp)
names(popular) <- c("country", "p1", "p2", "p3", "p4", "p5", "p6", "p7", "p8", "p9", "p10")

save(popular, file="popular.Rdata")

## countries and packages with downloads
cp <- vector("list", 221)
names(cp) <- countries$country
for(i in 1:length(countries$country)){
  message(i)
  a <- downloads[country==countries$country[i],]
  b <- arrange(as.data.frame(table(a$package)), desc(Freq))
  c <- as.character(head(b[,1],10))
  d <- head(b[,2], 10)
  cp[[i]] <- data.frame(package=c, downloads=d)
}

save(cp, file="cp.Rdata")

## countries data structure
country <- rep(names(cp), each=10)
rank <- rep(1:10, length(cp))
countries <- rbindlist(cp)
countries <- cbind(country, rank, countries)

save(countries, file="countries.Rdata")