By Suman Mukherjee and Forsyth Alexander
As the world waits anxiously for the fourth installment of the popular Jurassic Park movie series to be released Friday, we thought it would be fun to look into the social buzz for the upcoming summer blockbuster.
As fans, we were curious about things like, where the most Twitter chatter was happening, how tweets were breaking down by gender, overall sentiment, peak times for chatter, and more.
So we uploaded some Twitter data about Jurassic World into Watson Analytics, IBM’s natural-language cloud-based analytics service, and within minutes began unearthing pretty interesting insights, such as: the country with the most tweets so far is Chile; on the whole, women are tweeting more than men; and Portugal has the highest number of positive tweets, but also the most negative.
For each bit of research and each question we asked, Watson Analytics provided a visualization that summarized the information so we could understand what we were seeing quickly and easily.
That’s the great part about Watson Analytics – it’s all about opening up and extending the power of predictive analytics to people with little to no analytics experience to help them start making data-driven, strategic decisions.
Getting started is easy. After logging into Watson Analytics, we simply used the Twitter Data Connector in the service to import Twitter chatter for #jurassicworld from February 1 to May 29, 2015. As soon as we clicked on the imported Twitter dataset, the “guided” feature kicked-in and Watson Analytics shared exploration starting points. We didn’t have to type a word.
Based on the visualizations, we were able to see that:
- Although Chile had the highest total of authors tweeting about the movie, Mongolia wasn’t far behind.
- In April, women had more positive comments about the movie than in May.
- Overall, sentiment for the movie was positive, which was to be expected. However, there was no appreciable change in the order of sentiment over time. Positive was highest for each day, followed by neutral, then negative and then ambivalent.
- The global volume of all tweets (whether positive or negative) surged during the last five days of the month.
The surge of sentiment and tweets toward the end of May might be a reflection of the increased number of trailers being released, the Jurassic World tumblr page, and the fact that the release date was nearing. Regardless, such insights could be valuable for the later releases of the movie in other countries or could be used to drive earlier previews for more positive chatter sooner.
Adding language and city for even deeper insight into sentiment
We admit, we’re in the analytics business, but we were still fascinated by what we were learning about sentiment—and so fast. Now that we knew about which gender was tweeting more and when positive tweeting was at its highest, we once again took advantage of how Watson Analytics can process natural language by asking for sentiment about the movie by language, by cities and then by cities plus tweet type.
Watson Analytics answered our questions with scatter plots that showed:
- Only neutral sentiment was expressed in Arabic and in Ukrainian.
- The difference between positive sentiment and neutral sentiment in Norwegian was much higher than for other languages and the incidence of negative sentiment was quite low.
- Negative sentiment was higher in most of the cities I looked at, with positive sentiment staying steady until it reached Los Angeles. Also, there was a huge spike in positive sentiment from authors in New Orleans.
- The majority of negative sentiments came from shares and not original tweets.
The great thing about Watson Analytics is that it is accessible to people with little to no analytics skills. So, for example, a worldwide marketer for the studio or distributor or even a star’s publicist could use this information to adjust promotions or campaigns in the countries where Arabic and Ukrainian are spoken.
With our curiosity sated, we returned to work confident that Watson Analytics is embodying the quintessence of self-service analytics. Now, we can’t wait for the movie.