Insights

Visualizing the #Oscars2016 conversation

By Miguel Rios
Friday, 26 February 2016

With the 88th Academy Awards approaching on Sunday, we wanted to look at how fans on Twitter have been talking about the nominees. We built a visualization of the national conversation — broken down by state — about each #Oscar nominee in six of the biggest categories. Take some time to explore it (go ahead, we’ll wait!) and then read on to learn how it came together.

This Tweet is unavailable
This Tweet is unavailable.

(click image to access the full interactive)

We first had to decide what time range to use for our dataset, which was complicated by the fact that the nominated films were released across an ~8 month period. Measuring the entire period (the window of Oscars eligibility) biased heavily towards movies released earlier in the year, as it allowed for more conversation to accrue (particularly around secondary moments like DVD releases). We analyzed Tweet mention patterns for the nominated films to get a sense of how the conversations unfold:

This Tweet is unavailable
This Tweet is unavailable.

From this, we settled on a -7 through +30 day time period to be set for each film based on its US release date. But in this scenario, we were then missing the Oscars-specific conversation that unfolded recently (the data also showed an “Oscars bump”), so we also added in Tweet totals for the time period since nominations were announced (Jan. 14 to Feb. 23, 2016).

Another complicating factor was that several of the films nominated for Best Picture have titles that could be used in many unrelated Twitter conversations. For example, take “Brooklyn”. If you see a Tweet that says “I love those accents in Brooklyn!”, is it from a brogue-loving moviegoer or an #OH-ing tourist? We searched for film titles that appeared in Tweets alongside terms that suggest the conversation is film-specific.

With those questions answered, my colleague Krist Wongsuphasawat (@kristw) and I began turning the large dataset into a user-friendly interactive dataviz. We opted for a tile grid map to allow for a better view of the smaller states and for better mobile viewing, especially to make it easier to tap on each state to explore its rich data (we’d seen slightly different grid map layouts from multiple publishers and adopted the layout from the New York Times based on a study comparing six different layouts).

Once we began mapping the data, we quickly realized that each category seemed to have a clear fan favorite — meaning that a visualization of the most-discussed nominee in each state would have little diversity and not tell the richest story in itself. We added in the drop-down menu to view the second, third, etc., ranked films. Another way to explore the map is by selecting a specific state and seeing its full ranking of the nominees in each category, allowing you to compare how any two states are talking about the films or stars.

This Tweet is unavailable
This Tweet is unavailable.

Here are some trends we noticed:

  • Spotlight (focusing on the Boston Globe) performs better in Massachusetts — and the surrounding states — than elsewhere nationally
  • In the Best Actress race, perennial nominee Jennifer Lawrence dominates conversation nationally, but there’s a battle for second place between Cate Blanchett and first-time nominee @BrieLarson
  • While Kate Winslet takes 49 states for the Supporting Actress race, Rooney Mara was most-discussed in Colorado, where her film “Carol” had its premiere

We hope you enjoy exploring it for yourself!

This Tweet is unavailable
This Tweet is unavailable.
@miguelrios

Miguel Rios

‎@miguelrios‎

Data Analytics, Twitter

Only on Twitter