Analyze This: Past Oscar Winners

Porter Thorndike's picture
 By | February 22, 2016
in InfoAssist, Business Analytics, business analytics, Business Intelligence, Data Visualization, data visualization, ESRI, InfoAssist+, Oscars
February 22, 2016

Photo attributed to Zennie Abraham.

My favorite movie of all time is Fast Five featuring Vin Diesel, Paul Walker and Duane “The Rock” Johnson. It’s a silly, entertaining action movie that has a love triangle of conflict between Diesel’s posse of car thieves, The Rock’s elite unit of fugitive hunters and a South American cartel.
 
It was not nominated for a single award. Given that my own personal taste in movies will never be able to predict a future Oscar winner, I have decided to fall back on the historical data. I reached out into some sketchy, dark recesses of the interweb and located data on past Oscar winners and nominees by genre, viewership and ad prices by year and the location of best foreign film winners. Though we can't predict the future, this blog post will summarize some key takeaways based on historical Oscar facts.
 
I have in my arsenal the snappiest analytical tool on the market, InfoAssist+. It allows me to upload, wrangle, join and blend those three different data sources into a single, clean view.  The first thing I did was paint a Choropleth map of the winners of the Best Foreign Film category. Using the clean, accurate and handsome ESRI maps layer, the visualization clearly showed two countries that have dominated this category: Italy and France.  
 
Next I moved to analyze the winners by genre. The ranked bar is the best way to communicate this, and drama’s came out on top followed by romance. Deep down in places that I don’t talk about at parties, I am a massive super fan of The Notebook - the ultimate drama in my eyes.  I used the color bucket to show on the heat scale whether there was a disparity in the number of nominations vs. winners. The distribution of the heat scale remained true, # of nominations is tightly correlated and is an excellent predictor of the winning genre. 
 
Lastly, I wanted to analyze the viewership by year. I wanted to see which years had the highest viewership and the highest ad prices. I have options here like a multi-split axis, but for communicating two metrics over a time series I really like using the bar chart. Very quickly I saw that viewership peaked in 1998 and has remained cool since. Adding the ad price metric to the color bucket showed that the highest ad prices did not correlate strongly with viewership. The highest ad prices were in 2008 and 2015, which had very low ratings.
 
 
After all this, I’m waiting for the call from Hollywood’s super agents to get my help on fixing this high ad price, low rating debacle. Fluorescent pink tuxedos, sleeveless tuxedo’s and acid wash denim tuxedos are my three pronged plan for an Oscar’s viewership bonanza. If that doesn’t work, I am willing to crash the red carpet in a one piece ski suit with fur moon boots and a Daft Punk helmet. 
 
A good data storyteller needs analytical skills and an analytical tool that moves as quickly as their brain does. Many would argue that my analytical skills are in doubt (my wife and kids come to mind) but the quality of my analytical tool will never be in question. Once I found the data this entire process took only 15 minutes.

Based on the findings in this analysis of past winners, we could reasonably predict that future Oscars will go to Italian dramas, with high ad sales and low viewership. Hold onto your seats and your golden statues - we'll find out on Sunday!