Sentiment Ranking Methodology


The AMSTS College Football Sentiment Ranking is an experimental index that attempts to measure overall fan opinion of a college football team throughout the season.

IMPORTANT NOTE: The Twitter sentiment is in regards to the *total population’s* opinion of a particular program, based on hashtags used. One of the most common questions I get is why Alabama isn’t higher.

Background: The AMSTS College Football Sentiment Rankings scour hundreds of school and program-specific hashtags to amass a collection of tweets that is then run through a language sentiment algorithm that classifies tweets as “Positive”, “Negative”, or “Neutral”. It’s broken up into a couple of components.

Net Opinion: This is the score on how positive the overall tweet population is. To over-simplify a bit, if there are 1000 tweets, 500 positive, 300 neutral, 200 negative, the score would be +300 for that component. In actuality, tweets get graded on a scale from entirely negative to entirely positive depending upon the language used, then that is summed to come to a final number that gives the directional degree of sentiment.

Fanbase Passion: This is an indicator of how strongly a fanbase feels about their team, both by the volume of their tweets over a defined time period as well as the degree of attitude (positive or negative) displayed within the tweets. So, for example, Charlotte, who has very little tweet volume, and when it does it tends to be metered language, will have a different passion score than Alabama, with its fervent fanbase (and copious amounts haters).

Frequently Asked Questions:

Why isn’t [X] rated higher/lower:
I don’t know man, check Twitter. If people are heaping praise on an individual player during a bad performance, it may trend their ranking the opposite direction as a function of the words being used in the Tweets.

What does this actually tell me?
Ultimately, I think this is a measure of how a team is performing against their fanbase’s expectations. A team like Clemson will have high expectations, and one would expect them to perform at about a balanced 33% positive – 33% neutral – 33% negative throughout the season if playing to their fans’ expectations. The expectations might be a lot higher than, say, Kansas. So one would actually expect Kansas to have an “easier” time beating expectations.