FYI.

This story is over 5 years old.

Stuff

The Guy Who Used to Predict the Hottest 100 Reflects on Being a Kill Joy

We checked in with Brisbane-based Nick Drewe, who used to crunch votes on social media to accurately call the song list.

Nick and I went out for coffee. He brought lots of tech stuff along. Image by the author.

You might remember that guy who got himself in trouble for predicting the Hottest 100 a few years back. That guy was a Brisbane-based coder named Nick Drewe, who crunched votes posted on social media to accurately call the song list. His version of the list was smugly titled the Warmest 100.

In the 2013 countdown, he nailed 92 of the 100 tracks, including the top spot. The next year was tougher because triple j cracked down on people sharing their votes in an attempt to quash predictions. Yet, still with only 1.7 percent of the total vote, Drewe was able to predict the number one song—'Riptide' by Vance Joy.

Advertisement

With the annual music countdown just around the corner, VICE sat down with the Drewe to talk about pissing off triple j, spoiler alerts, and why he decided to stop publishing his predictions.

VICE: Hi Drewe, how did you come up with this process to predict votes?
Nick Drewe: I was scrolling through Twitter one night, right before voting closed for the 2013 Hottest 100. A bunch of my friends were posting this link to the triple j website with a list of songs they'd voted for. So I thought, I wonder how many people are posting this? I did a quick Twitter search for the link and I got thousands of them.

My coding skills back then were pretty amateur but I knew enough about Twitter's API [the site's code] to wrangle together a script that grabbed each vote people posted on Twitter. After an hour I found about 2000 posts, which was 20,000 individual song votes.

Did you know how you'd turn this data into something interesting?
No, every stage was just let's see what can be done. When I was writing the codes it was an experiment the whole time. I shot out a couple of emails to some friends that I thought might be interested, one was a writer and the other worked in the music industry. My friend Tom Knox had the idea to publish it and make the songs playable on a site.

So speaking technically how did you actually do it?
Twitter and Instagram have public APIs which let anyone write codes that interacts with their platforms so it's quite easy to get data out of them. And that's what I was really interested in—how people share information online and what we can tell about people.

Advertisement

We had no expectations for it, we just made a list and put it out there. I thought it might get picked up by some music blogs and a few people might talk about it.

Obviously your list annoyed some people. Take me through what happened after it was published online.
I woke up the next morning and it was like on the front page of the Sydney Morning Herald website. There was an overwhelmingly positive response but there were obviously trolls. I remember one article titled, "Brisbane Nerd Ruins Australia Day."

Some of the sports betting websites had odds for what the top song would be on the Hottest 100, but because we had the list, they shut down those markets straight away. Then they opened up new ones to see how accurate our lists was. I didn't make any money from the whole thing.

How did triple j respond?
I think there was an interview right before Australia Day where they touched on our predictions in a way that didn't mention the list. But I got an email from the ABC a couple of months after the whole thing. They just wanted to chat and brainstorm ideas for making the website and the voting system more secure. But then we went and did it again next year.

Was it different the second year around?
Yeah it changed completely. That was the time when Instagram started to blow up. All of a sudden people were taking screenshots of who they'd voted for and posting it on their Instagram account. So I hopped on there, used the hashtag, and the result was still as accurate.

So if it was working out, why did you stop?
The whole thing was about curiosity and seeing if it could be done. After we could do it, it felt like we ticked a box. There's a lot of copycat lists doing what we did and it's still totally possible to do it this year.

Did you learn anything about Australia predicting their annual tradition?
People share what they've voted for because it's a humble brag. It's just another representation of who you want to be online, not always who you are.

Follow Mariam on Twitter.