Sentiment analysis of baseball stadium reviews

I was cleaning my inbox over the Christmas break, and saw a 2017 message from someone working at company called reviewtrackers. He wanted to draw my attention to some interesting work that he had done at the company, analysing 130,000 online reviews of major league baseball stadiums (since the message, he seems to have left the company and moved into academia). The report is available here.

The analysis of the reviews offered insights into how overall sentiment had changed:

How the various stadiums compared to each other over, as well as across several aspects of the experience of visiting the stadiums, such as: Food and Drink, Fan Experience, Family Experience or Facilities.

The automated sentiment analysis of online reviews could be more insightful for companies than the automated analysis of Twitter conversations. This is because online reviews are typically longer than tweets and, therefore, less prone to misclassification. Though, they still need to be treated with caution.

At the end of the day, these reviews only tell us what fans are talking about. And, because fans are only going to talk about something that is worth mentioning (because it was either disappointing or surprisingly good), it would be wrong to make decisions only based on these conversations. For instance, according to this exercise, people talk more positively about the BBQ or the popcorn than about the hotdogs. Does this mean that they don’t care about hotdogs? Should we stop offering hotdogs, and focus on popcorn and BBQ, instead?

Not necessarily. We need to cross this dataset with others, such as sales data, or interviews.

Maybe fans don’t talk about hotdogs because it is a staple of the stadium visit. It’s what we call an “hygiene factor”. Something expected, not worth talking about, as long as there are no problems with it. Like free sugar in a coffee shop.

Offer customers good hotdogs, meeting customer expectations, and they will not talk about it. Yet, let the standard of hotdogs slide, or stop offering this option, and customers are bound to voice their disappointment.

Still, this message was an interesting example of the value of analysing online conversations. Thank you to Max for drawing my attention to this report.