And the Oscar Goes to…Data Analytics?
The 89th Academy Awards ceremony concluded recently at the Dolby Theater in Hollywood. The charismatic Jimmy Kimmel hosted the event, which had an ending that was reminiscent of the Steve Harvey incident at the Miss Universe 2015 pageant.
After the preview of the nominations, Faye Dunaway and Warren Beatty announced La La Land as Best Picture. Soon after, the entire cast and crew of the film rushed onto the stage, celebrating and ready to deliver speeches.
The only problem, however, was that La La Land wasn’t the winner. Apparently, Beatty had the wrong envelope. The real winner was Moonlight, a drama film directed by Barry Jenkins. The error was rectified amid apologies and rather awkward “exchanges”. Pun intended.
The trend of analytics firms making predictions in the area of sports and entertainment is growing each year. Data analytics solutions are capable of predicting a lot of things; an embarrassing moment, however, isn’t one of them.
That said, several analytics startups and data companies, every now and then, post predictions for various events of public interest — games, elections, movie and music awards, and so on — mostly to have some fun aside from flaunting their skill with numbers.
In the case of the Oscars, in general, some of the information sources that analytics companies leverage include:
- Major award ceremonies (DGA, PGA, BAFTA, SAG, etc.)
- Critics choice
- Previous nominations and wins
- Genre and box office performance
- Personal details of the nominees (age, records, etc.)
- Social media (public sentiment data)
- Demographics of the Academy Award voters
The 2017 Oscars are now behind us with many diversely talented artists recognized and honored. Let’s take a closer look at how data companies fared at the Oscars with their predictions for various award categories.
Taykey, an Israeli technology company, made predictions for the Oscars based on conversations and sentiments on the social media. The company specializes in real-time sentiment analysis and tracks hundreds and millions of data points every day.
Amit Avner, CEO at Taykey, said that Leonardo DiCaprio had the most influence on Internet users when the 88th Academy Awards ceremony was around the corner. DiCaprio received positive responses in 80% of the digital chatter about him. The talented actor, who is reportedly going to produce the Captain Planet movie, went on to win the Best Actor award for Revenant. Also, Matt Damon, who lost in the same category, scored only 57% positive responses.
Taykey used its proprietary methods to determine that Andrew Garfield and Isabelle Huppert had the highest positive sentiment ratings at 86% and 76%, respectively. Emma Stone, on the other hand, had a score of 67%.
So, here’s what Taykey predicted for four major categories:
- Best Actor — Andrew Garfield
- Best Actress — Isabelle Huppert
- Best Supporting Actor — Mahershala Ali
- Best Supporting Actress — Viola Davis
Interestingly, only half of these predictions turned out to be true, as Casey Affleck and Emma Stone won Best Actor and Best Actress, respectively. I personally didn’t mind Affleck beating Garfield to bag Best Actor, given the tangled mess that was The Amazing Spider-Man 2. A true comic book fan never forgets. But I digress.
Disrupted Logic Interactive, a provider of big data and artificial intelligence solutions for mobile, used 17 years of data points from the Academy Awards’ data. These data points were fed into the company’s software, which then used proprietary probability and relationship algorithms to determine the winner of Best Picture.
The software, which reportedly has an accuracy rate of “over 100%”, predicted that 66% of the voters in the Academy (more than 6,000) would choose La La Land. And for a few moments, the world thought the movie had won, except it didn’t. Again, this is a result I didn’t mind because La La Land, despite having some breath-taking cinematography, almost put me to sleep in the theater. (It is what it is.)
Luminoso Technologies, an artificial intelligence startup focused on natural language processing, predicted that the Natalie Portman starrer Jackie would likely win Best Picture. The company, which used 84,058 IMDb reviews to make the prediction, reported that terms like ‘cinematography’, ‘visuals’, ‘stunning’, ‘experience’, and ‘masterpiece’ had a direct correlation with the movies that got the nomination nod. However, it was Moonlight that took the coveted Oscar statuette.
In conclusion, many data companies got it wrong (especially the Best Picture category) for this year’s edition of the Oscars. This can be attributed to the fact that it’s not always possible to account for certain critical factors such as the campaigns run by each film studio leading up to the event. While some factors can be taken into account, others are difficult to determine, let alone study and “learn” from. It is therefore difficult for any analytics software to get all the predictions right.
Now for this year’s verdict:
And the Oscar goes to…Data Analytics. Not. (See what I did there?)
***
This write-up first appeared on the BRIDGEi2i blog.
