Go to Top

Can Social Media Predict the Final Four Winner?

I’m running an experiment to see if the social internet is correct in their collective opinions about the Final Four. Are we to trust what the internet tells us? Afterall, it is the medium that made people think ‘Rick Rolling’, LOLcats and the You Tube series ‘Fred’ were cool. There’s a lot riding on this, internet, so tell me what I need to know!

For this blog, we compiled data from the last seven days, scanning blogs, videos, images, mainstream news, micromedia (twitter & friend feed), forums & forum replies and Facebook. A main limitation here is Facebook and Twitter profiles with privacy settings, and when doing sentiment analysis, we hide the neutral comments.

Let’s check out what the internet says…

1270241766756

Duke is winning in terms of number of posts, followed by Butler, Michigan State, and finally West Virginia. Duke is easily the favorite to win, being the only number 1 seed left in the tournament, and they are getting the most attention online.

Butler is getting a lot of attention also, receiving over 83,000 mentions in the last week. If you’re wondering why I chose the color pink to represent Butler, there are a few reasons, the first of which is Duke shares the same team colors, and gets preferred treatment since they are the higher seed. Also, since they are being called this years ‘Cinderella’ team, I thought Pink was fitting. Lastly, since I am a huge Michigan State fan, I thought I would emasculate the opponent just a little bit. As we’ll see soon, there is a lot of buzz surrounding the Butler Bulldogs right now, but much of the talk is about how improbable their run has been, and not people jumping on their bandwagon.

Michigan Sate and West Virginia both have about 40,000 mentions in the last week, far less than Duke and Butler, but total number of comments is not always the most important indicator, especially online where people can say anything they want, nasty or nice, while hiding behind a pseudonym. And lets face it, there is no lack of opinions in the internets. This brings us to sentiment analysis. Is the content of the comments good or bad?

Lets start with the first matchup on Saturday: Michigan State vs. Butler.

Sentiment Analysis: ButlerSentiment Analysis: Michigan State

Butler has just over twice as many mentions as Michigan State, but only 5,568 (48%) of these are marked as positive sentiment. In fact, there were actually slightly more negative sentiment posts than positive for Butler! They’re a Cinderella team thats had a great run, but people just don’t have faith in them, and rightfully so! Their next game is against Tom Izzo and the unstoppable Spartans, who’s sentiment analysis shows a completely different picture.

Although far fewer comments than Butler, 5,842 (64%) of the mentions for the MSU is positive. It’s close, but Michigan State wins the battle for most positive posts with 5% more than Butler.

The late game on Saturday is Duke vs. West Virginia.

Sentiment Analysis: West VirginiaSentiment Analysis: Duke

A 1-seed vs. 2-seed matchup. Being the state of my birth, I’ve become a West Virginia fanboy the last few weeks, but it seems very few others have joined me. The Mountaineers have, by far, the fewest positive sentiment mentions of any of the Final Four teams with 794 (52.1%). And then there is the mighty Blue Devils, with a ridiculous 10,349 (55%) positive sentiment mentions. Advantage Duke. Sorry West Virginia fans…

This brings us to a MSU vs. Duke final. Now there is two ways to interpret the data here. Duke has nearly twice as many positive sentiment mentions as Michigan State, but MSU has a higher percentage of positive remarks.

Lets take a look at the mediums which these positive comments are coming from:

Is there a certain media type we put more weight into than others? I’ve always thought of Twitter as a megaphone that allows people to shout fleeting comments. I literally sneeze in 140 characters. The point here is that micromedia is nice, but doesn’t take a lot of brain power or conviction to type anything. Forum replies, although with no character restrictions, are much the same way. People enjoy bickering and arguing online, and forums provide some sort of organizated platform for that to happen. I think blog posts are created with a little more thought and conviction. At the very least, it gives the space to form an opinion and back it up in a space that you own. Given this, lets see how the positive sentiment for MSU and Duke match up through blogs.

2,613 (41.3%) of MSU’s positive mentions come through blogs. Compare this to just 1,694 (15.8%) blog posts for Duke. Aha! The ‘higher value’ positive sentiment posts favor the Spartans. It’s probably worth mentioning that the bulk of Dukes positive mentions (67.8%) came in through micromedia, meaning if my assumption that Twitter posts were of lesser value, then my conclusion is very very wrong.

Positive Posts by Medium - Michigan State

Positive Posts by Medium - Duke

Although I want badly to make the data favor MSU, it would seem as if the internets favor Duke. In fact, MSU is broken. Raymar is playing with a missing tooth, Delvon Roe has the knees of a fat 80 year old, Lucas is out with a torn Achilles tendon, Allen has a torn ligament in his foot… they look like the team was mugged in the parking lot. But Tom Izzo still has the faith, and I am an Izzoist.

Alright Duke, It sounds like the championship is yours to lose…

I’m running an experiment to see if the social internet is correct in their collective opinions about the Final Four. Are we to trust what the internet tells us? Afterall, it is the medium that made people think ‘Rick Rolling’, LOLcats and the You Tube series ‘Fred’ were cool. There’s a lot riding on this, internet, so tell me what I need to know!
For this blog, we compiled data from the last seven days, scanning blogs, videos, images, mainstream news, micromedia (twitter & friend feed), forums & forum replies and Facebook. A main limitation here is Facebook and Twitter profiles with privacy settings, and when doing sentiment analysis, we hide the neutral comments.
Let’s check out what the internet says…
[ insert Posts by School ]
Duke is winning in terms of number of posts, followed by Butler, Michigan State, and finally West Virginia. Duke is easily the favorite to win, being the only number 1 seed left in the tournament, and they are getting the most attention online.
Butler is getting a lot of attention also, receiving over 83,000 mentions in the last week. If you’re wondering why I chose the color Pink to represent Butler, there are a few reasons, the first of which is Duke shares the same team colors, and gets preferred treatment since they are the higher seed. Also, since they are being called this years ‘Cinderella’ team, I thought Pink was fitting. Lastly, since I am a huge Michigan State fan, I thought I would emasculate the opponent just a little bit. As we’ll see soon, there is a lot of buzz surrounding the Butler Bulldogs right now, but much of the talk is about how improbable their run has been, and not people jumping on their bandwagon.
Michigan Sate and West Virginia both have about 40,000 mentions in the last week, far less than Duke and Butler, but total number of comments is not always the most important indicator, especially online where people can say anything they want, nasty or nice, while hiding behind a pseudonym, and lets face it, there is no lack of opinions in the internets. This brings us to sentiment analysis. Is the content of the comments good or bad?
Lets start with the first matchup on Saturday: Michigan State vs. Butler.
[ insert MSU_sentiment ]
[ insert Butler_sentiment ]
Butler has just over twice as many mentions as Michigan State, but only 5,568 (48%) of these are marked as positive sentiment. In fact, there were actually slightly more negative sentiment posts than positive for Butler! They’re a Cinderella team thats had a great run, but people just don’t have faith in them, and rightfully so! Their next game is against Tom Izzo and the unstoppable Spartans, who’s sentiment analysis shows a completely different picture.
Although far fewer comments than Butler, 5,842 (64%) of the mentions for the MSU is positive. It’s close, but Michigan State wins the battle for most positive posts with 5% more than Butler.
The late game on Saturday is Duke vs. West Virginia.
[ WV_sentiment ]
[ DUKE_sentiment ]
A 1-seed vs. 2-seed matchup. Being the state of my birth, I’ve jumped on the West Virginia bandwagon the last few weeks, but it seems very few others have joined me. The Mountaineers have, by far, the fewest positive sentiment mentions of any of the Final Four teams with 794 (52.1%). And then there is the mighty Blue Devils, with a ridiculous 10,349 (55%) positive sentiment mentions. Advantage Duke. Sorry West Virginia fans…
This gives us a MSU vs. Duke final. Now there is two ways to interpret the data here. Duke has nearly twice as many positive sentiment mentions as Michigan State, but MSU has a higher percentage of positive remarks.
Lets take a look at where the positive comments are coming from:
Is there a certain media type we put more weight into than others? I’ve always thought of Twitter as a megaphone that allows people to shout fleeting comments. I sneeze in 140 characters. The point here is that micromedia is nice, but doesn’t take a lot of brain power or conviction to type anything. Forum replies, although with no character restrictions, are much the same way. People enjoy bickering and arguing online, and forums provide some sort of organizated platform for that to happen. I think blog posts are created with a little more thought and conviction. At the very least, it gives the space to form an opinion and back it up in a space that you own. Given this, lets see how the positive sentiment for MSU and Duke match up through blogs.
2,613 (41.3%) of MSU’s positive mentions come through blogs. Compare this to just 1,694 (15.8%) blog posts for Duke. Aha! The ‘higher value’ positive sentiment posts favor the Spartans. It’s probably worth mentioning that the bulk of Dukes positive mentions (67.8%) came in through micromedia, meaning if my assumption that Twitter posts were of lesser value, then my conclusion is very very wrong.
[ insert positive sentiment by media type for MSU and DUKE ]
Although I want badly to make the data favor MSU, it would seem as if the internets favor Duke. In fact, MSU is broken. Raymar is playing with a missing tooth, Delvon Roe has the knees of a fat 80 year old, Lucas is out with a torn Achilles tendon, Allen has a torn ligament in his foot… they look like the team was mugged in the parking lot. But Tom Izzo still has the faith, and I will never doubt him.
Alright Duke, It sounds like the championship is yours to lose…
facebooktwittergoogle_plusredditpinterestlinkedinmailby feather
, , , , ,

One Response to "Can Social Media Predict the Final Four Winner?"

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>