How Griffith University data scientist Bela Stantic predicted the Coalition’s surprise election win
Article By: Gavin Fernando www.News.com.au
A data scientist who predicted against the odds that the Coalition would win the election has explained how the opinion polls got it so wrong.
It was supposed to be the Labor Party’s “unlosable” election.
The opinion polls consistently predicted Bill Shorten would win on Saturday. Political analysts were sure of it. In the wake of the party’s shocking loss, many on social media have railed against the results.
But one man is not the least bit surprised. For weeks, Griffith University data scientist Professor Bela Stantic predicted the Coalition would sweep the victory.
Crunching the numbers at the Gold Coast university’s Big Data and Smart Analytics lab, Prof Stantic threw millions of tweets through his programs to find out what people were thinking about and feeling this election.
Last week, he told news.com.au Scott Morrison would remain prime minister, with the Liberal Party narrowly holding on to power.
Since his prediction came true on Saturday night, everybody has been asking the same question: how did the opinion polls get it so wrong?
Polling companies are facing increasing scrutiny, having shown the Coalition trailing Labor in every major poll for the past three years.
Prof Stantic said there were a number of factors at play, including sample sizes, the unpredictability of mobile phones, and a fundamental difference in how upfront people choose to be on social media versus in a phone poll.
“Firstly, my samples are much bigger,” he told news.com.au “I collected about two million relevant tweets from about half a million accounts. It’s a very big sample size from all around Australia.”
Prof Stantic — who also successfully predicted Donald Trump’s election win and Brexit — said his research had shown 5 per cent of social media data equated to a 95 per cent accuracy, if you have enough data.
He said the two million tweets he analysed in just a few days made up a greater representation of people than polls of 1000 people. He estimated his tweets equated to half a million people.
He also said the shift in opinion polling from landlines to mobile phone can make for inaccuracies.
“These polls are actually using phone lines which are not landlines anymore. People on mobile phones move around, so it’s not a clear prediction.”
He also suggested people are more “honest” on social media or talking to their friends about their sentiments on elections, but might be more “hesitant” when it comes to a phone-based poll.
In his work, Prof Stantic harvests a broad range of social media activity including Instagram, Flickr and Chinese social media site Weibo. In the case of the Australian election, he focused solely on Twitter.
At the end of the day, Prof Stantic doesn’t believe people who understand big data were surprised by his findings. “People who know my work immediately know that I can predict these results.”
“I made many public statements and Keynotes in the last month saying there was no evidence that Labor could win.”
If you have any further interest in data analytics or how this can help your business please contact Cappello Rowe for further information.