UPDATE: Triple J’s Hottest 100 is underway! Keep up to date with live updates and milestones here.
In the last few years, spoiling the results of triple j’s annual Hottest 100 poll has become a practice synonymous with the traditional Australia Day countdown itself and now a new player in the spoilers game has revealed themselves.
Earlier today, the Twitter handle @socialhottest began tweeting their Hottest 100 predictions, from 100 to 1, based on data collated via social media. While it’s not a new phenomenon — Warmest 100 have been predicting the Hottest 100 for the last few years — Social Hottest 100 claim to have a pretty large sample size of around 2000 voters, with roughly 10 votes each.
Music Feeds caught up with Michael George, the 34-year-old former computer science student behind the Twitter account, to find out more about his process, how confident he is in the predictions, whether he counted his own votes and who he reckons Taylor Swift — should she be counted — might knock out of triple j’s Hottest 100 of 2014.
WARNING: SPOILERS AHEAD!
After tallying over 20,000 votes using social media we will release the expected Hottest 100 tomorrow morning on Twitter #hottest100 #jjj
— Social Hottest 100 (@socialhottest) January 19, 2015
MF: Tell us a bit more about who you are. How many worked on the project?
MG: My name is Michael George, am 34 and studied Computer Science. We have done this for a few years, generally for the purposes of punting on the result and we landed some nice collects each time. This year two of us worked on the project. We could’ve used some more as it got very time consuming toward the end when more voters were getting in at the last minute.
Music Feeds: Tell us a bit more about your methods in calculating the results. Did you allocate points for each vote, or how exactly did it work?
Michael George: It was quite simple really. We sourced votes using the hashtag #hottest100 through Instagram and Twitter. This predominantly came up with results of those who had screenshotted their list of 10 votes (in rare cases less than that) and for each song listed we gave one point to that song on our spreadsheet.
At the end after going through over 2,000 posts from individuals (so over 20,000 votes) we had a large spreadsheet with a final tally for each song. We used the sort function and voila, there was the 100 we published today.
MF: You said data was collated via people who shared their votes on social media. Were they chosen at random or was there a method behind that?
MG: The method was simply using the hashtag. We feel we covered almost all votes shared on Instagram and about half of those shared on Twitter for the duration of the time voting was open. There was no picking and choosing, if they shared who they voted for it was counted.
MF: Did you take sampling bias into account?
MG: No, this is purely just the numbers we collected. While we agree some songs on the day will do quite a bit better or worse than our data suggests, we didn’t want to make judgment calls on how these would go.
MF: Do you have a background in statistical analysis? Or is this a passion project?
A passion project, or a punting project perhaps?
MF: While we’re on that subject, what are your motivations for sharing your Hottest 100 predictions? How long did it take?
MG: We just wanted to share it this year so everyone else could take a look at the work we put into it. On average it took a few hours each day the voting was open, much less at the start, much more at the end!
Watch: Japanese Wallpaper – Between Friends
MF: Did the social media campaign behind getting Taylor Swift into the count affect your predictions?
MG: It certainly affected the numbers. She had only a handful of votes before the social media campaign, and wound up in the 100 after all was said and done. In saying that, it would be the song I have least confidence in finishing where we’ve predicted.
One website claims to have gathered a few thousand votes for it alone, which along with the other votes could propel it much higher than our prediction, but then rumours are out there that it will be disqualified due to commercial promotion or that JJJ just won’t leave it in!
MF: If Swift does make it into the count, can you predict who might just miss placing in the top 100?
MG: By our count it would be Between Friends by Japanese Wallpaper.
MF: There’s already a site, the Warmest 100, which has been doing this for the last few years. Is there a point of difference between what you’re doing here?
MG: I think the difference would be little. Not sure if they are doing it again but if they are I don’t think you’ll see songs 10 or 20 places different to our count, maybe more like 5-10m but it would depend upon how many votes they counted through.
MF: How confident are you in your predictions?
MG: Very confident about the business end. I can’t see the top two not being the top two, but Chet Faker wasn’t far behind Peking Duk for number one according to our data. After that it’s a big drop off to number three, almost half as many votes.
In general while maybe only a few songs will be in exactly the same position on our count and the actual Hottest 100, I would expect we would be pretty close across the board. There’s always some surprises though, Warmest 100 famously didn’t have Internet Friends by Knife Party a couple years ago and it came in at about 55. Were people scared to say they were voting for it via social media? Well if you know the song, then probably yes!
MF: Did you vote in the Hottest 100 yourself?
MG: I did, and come to think of it I don’t even know if I counted those votes in our tally! I hope some Melbourne bands like The Smith Street Band do well on the day.
——
See the Social Hottest 100’s full list of Hottest 100 predictions here. You can follow Social Hottest 100 on Twitter, where they’re still answering questions.
UPDATE: Triple J’s Hottest 100 is underway! Keep up to date with live updates and milestones here.
