In August, students from Yale's Center for Customer Insights presented their findings from a study that used various IBM Analytics tools to analyze social media data from NFL fans. This data was used to gauge fans’ perceptions of the league and popular topics surrounding it and to then use these insights to drive effective social media interactions.
The first topic addressed players kneeling during the national anthem. The team generated word clouds based around this conversation to track the relevance and prevalence of the controversy among conversations between NFL fans. By tracking this data week to week, the Center was able to determine that the topic began to lose conversational steam among fans around week five of the NFL season, suggesting it was just a fad of an issue. Nonetheless, the traditional news media discussed the issue in one out of every five articles, keeping the topic relevant and leading to a long-term impact on the perception of the NFL.
Next, the team analyzed data on the topic of Thursday Night Football games. The unofficial general consensus of the time was that these games were extremely unpopular, so Watson Analytics for Social Media was used to see if this was in fact true. Collected data suggested that sentiments were overwhelmingly negative as expected. Using Watson Explorer, a deeper dive into the data suggested that fans were most concerned with effects on player health and an overall shorter amount of time for teams to prepare for these games to deliver a quality viewing experience.
Next came the notorious and never-ending controversy of the “catch rule” in the NFL, which defines what specifically makes for a catch. Initial analysis suggested that the only significant spikes in the mention of the catch rule happened during two controversial games during the season and as a whole accounted for only 1.2% of Facebook chatter. This initial data suggested that the controversy of the catch rule was only of passing interest to NFL faithful. Further analysis with IBM Data Science Experience captured the whole of the social conversation instead of sifting to find isolated keywords. Using a contextual combined search of “catch,” “touchdown” and “call,” it was instead found that the catch rule was the most salient topic across the NFL Facebook page last season.
Finally, social media data was used to measure the sentiments of viewers on the influence of concussions on the NFL. “Protocol” (referring to the NFL’s concussion protocol for a player suspected of a head injury) was found to be a top used term across NFL fans' social media. So clearly, fans were interested in this topic, but how did they feel about it? The team used Watson Explorer to analyze organic conversation to pick up on sentiment, finding that two main themes drove sentiment on this topic – player health and game impact. For player health, some praised the NFL for trying to protect its players while others admonished it for not doing more. For game impact, fans celebrated the return of their favorite players from protocol while others complained that their favorite player was being kept off of the field. This mirrored a trend in social media conversations last season following the NFL’s tightening of rules meant to protect players from concussions. On NFL related social media and among sports news sources, there was a significant increase in people praising the NFL’s efforts while on teams’ Facebook pages, there was a major increase in negative sentiment as fans complained about missing out on seeing players perform.
To learn more about the Yale Center for Consumer Insights’ study, watch their presentation of their findings online.
This project was executed by Jen Fischer, Florian Kogler, Vera Lo, Fernando Ortega, and Reshmi Suresh under the guidance of Ricardo Balduino, IBM data scientist.
#GlobalAIandDataScience#GlobalDataScience