CFP likability
Tweets show fan favorites
Twitter users and members of the College Football Playoff committee share positive feelings about the national semifinalists.
An analysis by Block Six Analytics of
75,000 tweets from Sunday until Tuesday based on percentage of positives posts ranks the top four teams in the exact order as the playoff committee: Alabama with 65.57 percent positive tweets followed by Clemson (65.22), Notre Dame (60.31) and Oklahoma
(46.77). Those numbers also mirror expectations for the semifinals with Alabama and Clemson expected to advance to the national title game.
The four teams also follow the same order as the Amway Coaches Poll.
“Since the CFB playoff rankings were released on Sunday, favorability has increased considerably for the top three teams in the Coaches Poll,” said Adam Robbins, B6A’s head of corporate sales. “Ohio State and Georgia saw the largest favorability drop after not being selected for the playoffs.”
B6A also looked at the top seven teams in the coaches poll for the entire season, analyzing 1.25 million tweets, which puts the margin of error below 0.5 percent, according to Alex Cordover, head of data science for B6A.
That data put Central Florida at the top of the likability ratings, just ahead of Alabama. UCF just finished its second consecutive undefeated regular season but was not close to making the playoffs. Only 16.54 percent of tweets about the Golden Knights were negative.
Strangely, Clemson received the highest percentage of negative tweets for the season (28.80) of the seven schools but the least negative since the playoff pairings were announced (7.18).
“The number of impressions is only one component in evaluating social media conversation,” Robbins said. “Understanding sentiment and engagement is crucial in measuring social media value, especially related to sports. It allows teams, leagues and brands the ability to identify topics, trends and people that are generating the most interest and value for their organization.”
Favorability Ranking Methodology
Sentiment analysis is a natural language processing method designed to analyze the emotional polarity of posts, comments and conversations. Sentiment has two components: a classification into “positive,” “negative,” or “neutral” and a determination of how strong the emotion is. To estimate sentiment, tweets are broken into their component words. Each word has a score which describes how strongly it contributes to positivity or negativity. The scores for the tweet are summed together and normalized giving the final sentiment magnitude. If the magnitude is less than 0, it is a negative sentiment, and if it is greater than 0, the sentiment is positive. Since the magnitude of sentiment directly affects marketing lift, B6A’s sentiment methodology is uniquely suited to analyze social media conversations.
B6A Description
Block Six Analytics (B6A) is a sponsorship analytics company whose clients include Fortune 500 brands, large marketing agencies, esports organizations, and teams in the NFL, NBA, and MLB.
Our Partnership Scoreboard enables clients to leverage artificial intelligence, computer vision and natural language processing to determine the crosschannel value of sponsorship campaigns in near-real time.