Uncommon Sense: Batman v. Superman

Finally, we can settle these age-old questions once and for all:

Behold! TweetSense instantly reports on how twitter users feel about what they tweet. For example, if someone says:

  • 	"Hated the new Man of Steel movie, it sucked #superman."

This tweet would bring superman's sentiment rating down. However, let's assume a tweet says:

  • 	"Loved The Dark Knight Rises, it was Spectactular #batman".

Then Batman's user sentiment would go up.

After analyzing hundreds of tweets of the user's choosing from the last seven days, a wonderful graph shows up with the overall user sentiment rating.

Behind The Scenes

We're proudest of our semantic tweet analysis (the engine that reads each tweet and is able to determine how the user felt) and equally so about locally-weighted linear regression used to predict future trends, all written from scratch.

Our code parses through hundreds of recent user tweets upon request (obtained through Twitter's API) and analyzes each tweet's sentiment regarding a specified term. It then displays wonderful line graphs thanks to Google's Chart API.

It's amazing how natural the resulting extrapolations turn out! They reflect true changes of heart. See for yourself how our predictions of, for example, "MHacks", compare to reality! Hint: they're pretty accurate.

We mined our source data from Twitter's awesome API. We used Heroku as our web host, Python for our backend, while Javascript helped translate numbers into graphs. Of course we used responsive HTML5 and CSS3 for this super cool website. In the future, we hope to integrate Alchemy's API (when we can afford the calls) and Bloomberg's API in the long term to provide even more interesting data.

The Team (Waffles@1)

Peter Mauldin (click to collapse)

Worked on the Back-end and Front-end.

Helped put together data from the API's and scripts on the webpages to make it all work

contact at: petermauldin@utexas.edu

William Widjaja

Worked on the scripting, design, and graphical output

Wrote the algorithm to parse and graph the back end data. Also built the HTML and responsivity

contact at: wwidjaja@imsa.edu

Samuel Tenka

Math genius. Wrote Gaussian distribution weighted average methods from scratch to predict future sentiment

Converted data into semantic information which could be understood by humans.

contact at: samtenka@umich.edu

ByungHoon Lim

The first line of contact with the Twitter and Bloomberg API's

Parsed seemingly arbitrary data into usable data on the backend

contact at: seian.hoon@gmail.com

Why "Waffles at 1"?

On the first night of MHacks V, students were scheduled to have a waffle break at 1 am

In a dramatic turn of events, the waffle machine blew a fuse, and thus no waffles could be consumed by the zealous Hackers. The team then decided that the waffles were so awesome, that they wrecked the waffle machine. We used that energy to produce the awesomeness in front of you. And will continue to carry on the legacy of the late ... waffles@1