Baseball is back, and fans came out of months of hibernation to watch their favorite team play their first game of the year. Social media, particularly Twitter, allowed fans to share their experiences, reactions and news in real time.
I was able to sample over a half-million tweets predominately from fans containing any of the 30 Major League Baseball teams’ mascot names or Twitter handles. From this I was able to construct time-series graphs reflecting the Twitter volume at any given time over the course of the two days of Opening Night and Opening Day games.
Below is a composite chart of Sunday through Monday night delineated by team. Typically, each team’s heaviest traffic comes during their first game. However, San Diego and Atlanta bucked this trend when the Craig Kimbrel and Melvin Upton trade was announced. In fact, this news created the highest tweet volume of the entire time frame.
To enhance the snapshot of the tweet activity, I attached a sentiment score to each tweet and averaged it for each 5 minute collection bin. The scores range from -1, a strongly negative tweet, to +1, a strongly positive tweet, and that is denoted using red and green, respectively. Generally, most tweets were positive. Intuitively, you would expect this since fans are generally upbeat and optimistic about their team’s season or for the sheer fact that the baseball season has started. There were definitely a few instances when the tweets become rather negative. For example, the Cubs had bunches of negative tweets after their game ended most likely due to negative press regarding the renovations at Wrigley Field.
I’ll let you explore the data visualization instead of pointing out every instance of negative fan reaction. There are two ways to view the data: picking two teams to compare or picking one of the 15 games. Please note that this is toggled by the buttons at the top of the visualization, and you must also select the date you want to view. The date should default to Sunday to show the Cardinals-Cubs game, but if you would like to examine other games, you have to switch the date to Monday.
On the lighter side of social media analytics, I did a quick analysis on the emoji used by each team’s fans. Below is a table of the top 5 emoji that fans of each team used over the course of the two days. The baseball emoji was clearly the most popular (and obvious) emoji for this occasion, and most of the other emoji reflect the team’s colors, mascot, or mood (check out the Braves and Phillies). The Brewers had a rather interesting emoji turn up in their top 5: a basketball, which reflected the University of Wisconsin playing in the NCAA Men’s Basketball Championship Game last night.
Notes: Twitter allows me to only sample a real time stream of tweets; this data set does not represent the entire volume present throughout the time frame. I also filtered out tweets by looking for precise Twitter handles including the ‘@’ sign or the presence of ‘opening day’ or ‘opening night’. This mitigated the problem of scraping tweets about generic athletics, angels, padres, or pirates.
Sentiment analysis was conducted using the TextBlob package for Python.
I code a bunch of things here. I really need to update my blog about statistics at stats.seandolinar.com.