In a story that might be one of the biggest to hit baseball going forward, the New York Mets were recently sold to hedge fund owner, Steve Cohen, for a record $2.475 Billion. Cohen, a lifelong Mets fan himself and a true New Yorker down to his accent, immediately became not just the richest owner in baseball, but the second richest owner in all of sports. He is basically living every little kid's fantasy to buy his favorite team and run it with the sole intention of steamrolling every championship. As a lifelong Mets fan myself, I could not be happier than to let him do just that! The question I had though was how are other Mets fans reacting to the news?
I was able to scrape Twitter data user the rtweet package in R. I had to make a few compromises for this analysis. While I wanted a true grab bag of random Mets fans, this was ultimately not realistic given the limitations of the package. While I could search for all tweets by "#Mets", it would only go back five days. However, it allows you to get up to the most recent 3,200 tweets of any one user. So instead I selected eight Mets beat writers and bloggers who were active on Twitter this year to serve as my proxy. To give them their proper kudos, they were:
I only included original tweets (no retweets) in which they directly mentioned the word "Mets" or named one of the key figures in Mets management, which provided 9,464 tweets and 234,618 words (24.8 words/tweet). However, I then filtered out all stop words (common words or non sequiturs that have no meaning in Natural Language Processing), as well as all hyperlinks, direct mentions of a Twitter username, hashtags, numbers, and dollar amounts. This brought it down to 83,405 words (which means 64.5% of words were removed).
Here are a list of words most mentioned by Mets Twitter this year, not including any player or personnel names (as well as filtering out the word "Mets"):
A lot of common baseball words for the most part ("game", "baseball", "team", etc.). A few that stood out were "owner" and "gm" (both positions changed in 2020), "Yankees" (the oil to the Mets vinegar), "RHP" (the Mets were notably lacking left-handed starting pitching this season), and "fans" (this season of course had none).
The calculus changes when we include proper names. I tried my best to group similar references to the same person together. Let me tell you, this was quite the challenge! Some names were easy to group (Luis Rojas was usually referred to as either "Luis", "Rojas", or "Luis Rojas"), but then you had the players with nicknames. Mets first baseman Dom Smith could sometimes go by "Dom" or "Dominic". Or former Mets GM, Brodie Van Wagenen, who might go by "Brodie", "Brody", or even just "BVW". But nothing (AND I TRULY MEAN NOTHING) was worse than trying to group all of the nicknames for Alex Rodriguez and Jennifer Lopez -- "A-Rod", "Arod", "J-Lo", "J-Rod", and on and on all 14 of their references went. All of that effort and yet they only had 66 mentions, so they were far lower on this list than I assumed they would have been. Good thing they didn't actually get the team!
No surprise that the sale of the Mets dominated the news cycle this year. Cohen was by far the most referenced word (1,101 mentions) with his appointed team president, Sandy Alderson, number three (672). I did find it funny that unpopular former owner, Fred Wilpon, had 60 percent fewer mentions (443) as the more revered Cohen, despite the fact that I actually combined all his references with his son, Jeff's (it was too hard to distinguish if a mention of "Wilpon" referred to either individually, so I had to combine them into "fjwilpon"). Cy Young contender, Jacob Degrom; breakout players, Michael Conforto and Smith; and Robinson Cano, who recently was suspended for 2021, were also top 10 in mentions.
Here they are as word clouds:
I then ran a sentiment analysis on the tweets. Sentiment analysis is an indicator of how positive or negative the connotation of language is. I wanted to group all tweets by month written from March through December (or at least December 16th when I wrote this) to see if the sale to Cohen had any positive impact on sentiment. The data clearly showed "The Cohen Effect" by the end of the year.
Cohen entered into exclusive negotiations to buy the team on August 28th and was approved by the MLB owners on November 2nd. While September was likely dragged down by the Mets finishing the season out of the playoff run, it did see an immediate uptick in sentiment and positive tweet percentage over the summer (naturally, sentiment was low when the season was ongoing and the Mets underperformed expectations). By the time the season ended in October, the Mets news cycle had turned almost exclusively to the pending sale, which caused the positive sentiment rating to double over September and triple July. Not only was sentiment among the highest of the year, but the three months since the season ended have also seen over 70% positive tweets. The only other month to see similar sentiment and positivity was June, which was buoyed by the optimism regarding the announcement of the 2020 baseball season returning after it had been delayed back in March.
Here are the average sentiment by player/personnel name, sorted by the most and least positive.
Notice a trend with all the positives? The sentiment analysis seems to show that fans romanticize about the players they do not have more than they appreciate the ones they already do. George Springer, Trevor Bauer, JT Realmuto, Francisco Lindor, Nolan Arenado, and James McCann were all big name free agents or trade targets that Mets fans were excited at the prospect of the team trying to acquire (the Mets have since signed McCann). Also notice overwhelmingly positive sentiment for Cohen, Alderson, and their new GM, Jared Porter.
On the negative side, most of these make sense. Jed Lowrie recently had his contract expire after never recording a single hit in two years as a Met, Billy Hamilton was part of an in-season trade by the previous regime that was generally maligned, and Michael Wacha was another free agent bust. Jeff McNeil is an All Star player for the Mets, but he did start last season in a slump that likely got a lot more negative attention than his good games. However, I have no idea what poor Luis Guillorme did to get on Twitter's bad side. In fact, Guillorme hit a career high .333 last season, so I am genuinely confused. I would hypothesize that he may have been paired in some tweets with Lowrie, whom he replaced, but I have not done any further analysis on pairs of words as of this time.
Finally, here are the least and most positive associated words with Mets tweets this year:
So are Mets fan excited about Steve Cohen buying the Mets? It sure seems that sentiment is true from analyzing their tweets. However, this is still technically the honeymoon phase. I look forward to revisiting this topic in the future as the team either meets or fails to meet the aggressive mission statement he put out.
Comments