apply()-woes.r

library(dplyr)
library(data.table)
library(xlsx)

# Define your workspace: "X:/xxx/"
wd <- "D:/gitlab/analytics/"
setwd(wd)


f.drop.players <- function(server.size = 1000, playtime = 90, dispersion = 0.7) {
  t <- sort(rlnorm(server.size, log(playtime), dispersion))
  result <- data.frame(playtime = playtime, 
                       dispersion = dispersion, 
                       fully.empty = t[server.size]/60, 
                       percent90.empty = t[ceiling(server.size*0.9)]/60)
  return(result)
}


# playtime -- sequences of 15mins
x <- seq(45, 120, by = 15)

# run each draw 100 times
times <- 100
result <- vector("list", times)
for (i in seq_len(times)) {
  # this generates a list of lists: as.data.frame(t(mapply(f.drop.players, playtime = x)))
  result[[i]] <- do.call(rbind, lapply(x,function(x){f.drop.players(playtime = x)}))
}
# bind into a single dataframe
result <- rbindlist(result)

# get results
dta <- result %>% 
  group_by(playtime, dispersion) %>% 
  summarise_each(
    funs(mean, median, min, max)
  ) %>% 
  select(
    playtime, dispersion, starts_with("fully"), starts_with("percent90")
  )

# output to excel
write.xlsx(t(as.data.frame(dta)), "server-spindown.xlsx", col.names = FALSE)