Are you a Bachelor(ette) superfan like James S. or just easily amused like me? Either way, you’ve probably wondered what the rest of the world is thinking about your favorite (or least favorite) show contestants. Using Watson Analytics for Social Media—a nifty add-on for the ever powerful self-service analytics tool from IBM—I was able to do just that and get a statistical perspective on how the Twittersphere really feels about the Bachelorette Season 12 contestants in just about 5 minutes.
First, I had to tell Watson Analytics what I wanted to do
The first step in the process was choosing the contestants I wanted to analyze. I told Watson Analytics to pull social media data related to the topics in the image only. Since names like Chad, Alex, Jordan and Wells are not unique to the show, I told Watson Analytics only to include the topics when related to keywords like “bachelorette” and “@BacheloretteABC”:
Once I had them all entered I moved on to some themes that would be telling of peoples emotions towards a particular character. I added superficial words like “hot,” “love,” “hate,” “favorite” and “win.” You’re welcome.
I chose to analyze social media data from Monday, May 24th (the day the show opened on air) up to June 2, the day before I wrote a blog on this topic for LinkedIn. I analyzed data in English, French, German and Spanish, and chose to look at social media data from Twitter, Facebook, blogs, reviews, videos and forums.
Okay, now for the fun stuff—the results!
As you can see from tree map at the bottom of the image above, Chad dominates the total share of mentions to date, followed by Jordan, Alex, and Daniel. However, just looking at week 1, it was all about suave, confident, former pro quarterback, brother of Aaron Rodgers, Jordan. That same week though, we had a lot of talk about the Canadian, Daniel, which makes a lot of sense since he decided to strip out of his clothes for a couple laps around the pool with absolutely no help from Uncle Jack Daniels.
Week 2 was something of a different story, though, because we started to see Chad’s true colors. He wasted no time stirring things up in the house with his strong alpha male personality, brutal “honesty” and impressive eating habits.
Showing off the Watson name
It’s great to see the trend of a topic over time, but that doesn’t necessarily tell us anything about the emotions expressed around those topics. This is where Watson Analytics shows off where it got its name from. Big daddy Watson won Jeopardy! in 2011 using super computer strength, machine learning and natural language processing capabilities (among others), and IBM has built some of those functions into Watson Analytics so it can understand human language and attach emotion to sentences. That allows Watson Analytics to score the sentiment around a specific topic, such as being positive or negative in the simplest form.
From this dashboard, we see that of all the mentions about Chad and Daniel, more than half are negative. Ouch! This should not be that surprising for anyone that watched the first two weeks of the show, considering their onscreen antics.
The next thing I investigated was the breakdown of conversations about the themes I chose at the beginning (see the following visual). From the visual on the left, it looks like Bachelorette fans are more apt to discuss their love for contestants than any other theme I chose, followed by who they think will or want to win.
On the left, we see the breakdown of theme mentions per individual topic (contestant name). It seems that people have a love-hate relationship with Chad, because people chose to use the word love more for Chad than anyone else even though their overall sentiment for the poor fella was negative. Perhaps people are tweeting things like, “I love Chads honesty, but he’s a real jerk,” because he is also the most “hated.” On a separate note, it looks like people want or think they know Jordan is going to win. (I mean when you know Aaron Rodgers is coming over for Thanksgiving turkey, how could you think about picking anybody else. Amirite?)
In a New York minute
Two other interesting visuals I’ll show are the gender breakdown and the mentions by geography. Apparently men are talking about Christian the most.
Not sure why men like Christian, but the show has taken place in California so far and JoJo (the Bachelorette) is from Texas. New Yorkers just love drama, obviously!
I can only talk about the Bachelorette with a serious face for so long…
What I wanted to show was how easy it is to gain valuable insights into social media topics with Watson Analytics and to show the importance of looking at multiple types of analytics when analyzing projects. For instance, if we were just looking at trending topics over time, we might not know how people actually feel about certain topics, or if we just looked at theme mentions per topic without looking at topic sentiment, we might think everyone loves Chad the most.
Let me know what you think about the Bachelorette Social Media analytics in the comments. Do you agree with the public opinion on your favorite or least favorite contestants?
If you found the social media add-on for Watson Analytics interesting, you can try it for free at the Watson Analytics for Social Media website.
Or if you just want to use Watson Analytics, it can be used for free in perpetuity by signing up at the Watson Analytics homepage.
And, happy binge watching!
About the blogger
When he’s not kicking back to watch reality TV or or barreling down a Red Bull Crashed Ice track, Kirk Thornton is an analytics solutions client representative for IBM, helping organizations solve their big data challenges with analytics. His specialty is Watson Analytics and his personal mission is to eradicate the barrier to entry for organizations and to remove the stigma that analytics tools are complicated.