AI Safety

NFL tweet study reveals fan sentiment drops at kickoff, diverges by halftime

Researchers mapped 4 seasons of NFL tweets to uncover predictable emotional arcs in fandom.

Deep Dive

Researchers from the University of Vermont's Computational Story Lab analyzed over four seasons of NFL game-related tweets (2011–2014) to quantify how team performance and geography shape fan sentiment. They found a remarkably consistent emotional trajectory: pre-game sentiment is positive for both fan bases, but it drops sharply after kickoff. By halftime, sentiment partially recovers, but the second half reveals a stark divergence—fans of winning teams experience a sentiment surge toward the end of the game, while losing team fans' sentiment remains depressed, ending well below the starting point.

The team also introduced a 'fandom radius' metric, identifying geographic regions where engagement significantly exceeds baseline discussion levels, effectively mapping each franchise's local fan base. A comparison between sentiment and win percentage revealed only a weak positive relationship, indicating that social media happiness depends on factors beyond simple wins and losses—such as rivalries, game context, and individual player narratives. This work contributes to computational social science by providing a quantitative framework for understanding modern sports fandom.

Key Points
  • Sentiment for both winning and losing teams drops at kickoff and partially rebounds at halftime.
  • Post-halftime sentiment diverges: winning team fans see a late-game rise; losing team fans stay low.
  • A weak positive correlation was found between team win percentage and overall sentiment, suggesting other factors matter.

Why It Matters

Quantifies the real-time emotional rollercoaster of sports fandom, with implications for social media analytics and fan engagement strategies.

📬 Get the top 10 AI stories daily