In the last couple of years, the mystique and mystery of Aussie youth radio station Triple J’s Hottest 100 has been dampened somewhat by the easy availability of voter statistics — Facebook posts, Instagram screenshots and Twitter lists of listeners’ top 10 music tracks of the year. In 2013 and 2014, The Warmest 100 used stats to create an unofficial playlist of the top 100 songs, and this year it’s the Tepid 100 that thinks it’s on the money.
The Tepid 100 is a project by Uni of Melbourne student Ed Pitt, inspired by the data-mining success of the Warmest 100. While the Warmest 100 ran into a bit of trouble with Triple J changing the format of its votes on social media, Pitt’s work is a lot more low tech than the optical character recognition of previous years — it just required a lot more hard work.
Manually entering around 20,000 votes from the Hottest 100 hashtag on Instagram into a series of spreadsheets was the time-consuming part of creating the Tepid 100; not nearly as much effort went into building the website itself, based on a generic Tumblr layout. While the Warmest 100 collated nearly 1800 votes and 1.3 per cent of the total 2013 voting spread, this year’s data mine of 2064 ballots is likely to be a slightly higher percentage.
The sample was taken from the starting of voting until 6:55PM on January 7 — a fortnight before the close of voting in one hour from now, at midday on the 22nd. It’s possible that early votes could have a different trend to later votes, but Ed is confident of the accuracy of his predictions. If you want to see what the Tepid 100 predicts to be the top 100 and 101-200 songs in this year’s Hottest 100 countdown, click here and here. For a look into the data science behind the Warmest 100, listen to the episode of Download This Show below. [Tepid 100]