Patiently waiting so that one day I can brag to my data friends I got a shirt from Rstudio hah
2 Likes
It's all about the free t-shirts.
1 Like
)Sorry for the missed week (#8). This week we'll randomly select 10 winners to catch up!
3 Likes
Oh there'll be 10 winners this week? As an Italian I can only say...
2 Likes
3 Likes
And the winners are...
1 Like
We missed a week! Week #8 and #9 winners will receive an RStudio t-shirt and all of the hex and RStudio stickers we can find. We will also have t-shirts and stickers at rstudio::conf! Time to register is running out and we expect to sell out the entire conference - so act soon! The winners for the week of November 10th & 17th are:
| Name | likes |
|---|---|
| jakekaupp | 12 |
| taraas | 10 |
| hinkelman | 5 |
| tyluRp | 5 |
| Tazinho | 3 |
| Anantadinath | 2 |
| scw | 2 |
| raviolli77 | 1 |
| yonicd | 1 |
| yutannihilation | 1 |
@jakekaupp, @taras, @hinkelman, @tyluRp, @Tazinho, @Anantadinath, @scw, @raviolli77, @yonicd, @yutannihilation
7 Likes
Yes finally!!
Posting a meme I had created for my organization to show my gratitude towards R. Hope admins don't mind
1 Like
Someone has to be at the far end of the probability curve.
If you want to check how unlucky you are, relative to other people eligible to win (having likes), here is some code (it does make a few assumptions that the relative probabilities between people have been similar in previous weeks, but I do not otherwise have the data on that one, also a few assumptions in the way the probabilities change with each winner). Starting from the EconomiCurtis users data set creation earlier in the thread after the command
users = users %>%
distinct(name, .keep_all = TRUE) %>%
mutate(
DateTime = Sys.time()
)
It takes that users dataset and calculates the chance that people (with 1 or more likes) who have not won have had of winning
iAm <- "thoughtfulnz"
library(dplyr)
library(ggplot2)
# past winners
ordered_winners <- c("mara", "alistaire", "daattali", "emilyriederer",
"eric_bickel", "nick", "jessemaegan", "raybuhr", "billR", "mmuurr",
"apreshill", "pavopax", "mfherman", "rensa", "tomtec", "pssguy",
"cdr6934", "rpodcast", "Andrea", "timpe", "cderv", "PirateGrunt",
"mungojam", "rkahne", "DaveRGP", "danr", "jim89", "nakamichi",
"Ranae", "jasonserviss", "FlorianGD", "Damian", "jtr13",
"RStudioUser", "robertmitchellv", "jakekaupp", "taraas",
"hinkelman", "tyluRp", "Tazinho", "Anantadinath", "scw",
"raviolli77", "yonicd", "yutannihilation")
#exclude employees
users <- users %>% filter(title != "RStudio Employee" | is.na(title))
draw_odds <- function(x, winners, drawees){
#zero out likes to represent having won so not getting picked again
drawees$likes[drawees$name %in% winners[(1:x) -1]] <- 0
#figure probs
probs <- data.frame(pvl = 1 - drawees$likes/sum(drawees$likes))
#leading zeros on asusmption I may be unlucky for some while
names(probs) <- paste0("draw", formatC(x, width = 4, format = "d", flag = "0"))
return(probs)
}
# probs for each draw of not winning
the_odds <- bind_cols(lapply(1:length(ordered_winners),draw_odds,
winners=ordered_winners,
drawees=users))
cummlative_prob <- 1- apply(the_odds, 1, prod)
results <- data.frame(user = users$name, cummlative_prob)
results$cummlative_prob[results$user %in% ordered_winners] <- 0
candidates <- results[results$cummlative_prob > 0,]
candidates$who <- "everyone else"
candidates$who[candidates$user == iAm] <- "me"
ggplot(candidates, aes(x=cummlative_prob, colour=who, fill=who)) +
geom_dotplot(dotsize=0.3) + facet_wrap(~ who, nrow=2) +
ggtitle("Unluckiness of non-winners") +
xlab("unluck (cummulative chance of winning, but didn't)")
6 Likes
Thanks for an awesome place to talk about R! Hopefully one day I'll make it to rstudio::conf, hard to swing west coast travel. Hopefully it gets big enough that you decide to move east some year!
3 Likes
This is the best thing I've seen all month, no lie.
1 Like
Any Week #10 winners? 
I want to float the idea (and I figure I have the moral mandate to do so as I haven't won a T-shirt), that as a long term sustainable encouragement add a total likes threshold- However often it gets run as a process, if total likes among eligible candidates is above x, then that triggers a draw.
This would both make it a long term encouragement to participate, and keep the process transparent for people to check themselves.
The difficultly with time base draws is it is predicated on an assumption of activity in that time.
3 Likes
I like this idea, I haven't won either but have a decent chunk of likes overall.
1 Like
Thanks for the stickers 
Also, use a credit card to iron out air bubbles. I almost had a mini heart attack when messing up the dplyr sticker.
4 Likes
How do these stickers peel off, BTW? In case I need to remove them later?
1 Like
You just get a new laptop 
3 Likes