To the Twitter Tune

  We know how popular it is to take the tweets of a certain exceptionally active twitter fanatic and attempt to use some method to create pseudo-random text. It is funny and almost always nonsensical and can lead to a good bit of fun and time wasting when complete. But that's been done before. Now, our group wanted to see how much time could really be wasted if you take a massive set of tweets from your favorite Twitter personalities and put them to the tune of your favorite song. Imagine the possibilities! Closer by the Chainsmokers as written by Donald J Trump, World Star Hip Hop bringing you their version of Shape of You by Ed Sheeran, all the inner Kermit the frog memes you can find replacing the masterpiece that is Stitches by Shawn Mendes. 
 First, you are asked to input a song, its artist, and the twitter personality you would like to see. Up to the most recent six thousand tweets are retrieved, then cached. Using the Nltk API, our program goes through both the tweets and the song lyrics, taking all of the nouns and adjectives, and giving each word a syllable value. This way, the rhythm of the song is kept largely intact and the sentence structure is similarly preserved. It then presents the original, and remixed version for you to enjoy.
 At the beginning, our idea was to make use of Markov chains to maintain some sort of syntactical sense, but using a singe replacement Markov chain yielded only gibberish. In lieu of using a double or triple replacement Markov chain, the decision was made to instead make use of the Nltk API and assign parts of speech to available words. 
 As a group, we're extremely proud of the fact that our program works at all. Only one of us had done a hackathon in the past and the majority of our group had never even coded in python. The amount of libraries used and files integrated may not truly be that large compared to others, but it is something we take great pride in. It was a true learning experience, and that might be the best part. Those who already knew python learned a great deal and those who didn't know python to begin with truly excelled despite a lack of experience. There may not be too many places to go with what might be considered a "silly" program, but we are incredibly excited with getting it to work under the time constraints and having fun while doing it.

Built With

• amazon-web-services
• bootstrap
• flask
• github
• javascript
• python