Inspiration - We wanted to create something whimsical and humorous using the Twitter posts and quotes of others (who are either fictional, or too old or dead to have a twitter account)
What it does - Twitter API to collect the most recent 200 tweets of Jaden Smith, and then concatenates them and puts them into a JavaScript version of a Markov Model. The user of our website then starts a sentence in the text box and clicks "Speak!" to use the Markov chain to develop a quasi-random most-likely end of the sentence - by Jaden himself (his last 200 tweets). The last part of the user's input is appended to the front of the text before Markov rendering to ensure that the user's start of the sentence is the start of the Markov generation (the output). The user can select other premade options (namely the Terminator, Mahatma Gandhi, Nicki Minaj, and Rick from Rick and Morty) which use large text files we created of quotes of those people/characters to create a Markov chain in the same manner, enabling the user to enter the beginning of a sentence and have it completed using Markov prediction by that person/character. The user can select different people to create a mish-mash story of sorts, beginning sentences in a humorous way and having first Gandhi finish the sentence, then Nicki, then the Terminator, all in one "paragraph." The user can switch between people and senctence-starters at any time during the creation of the paragraph. At any time the user can click the "Clear" button to start a new paragraph/story afresh.
How I built it - See above for description of Markov. First we created a front-end website and simultaneously created the Markov generator. We then utilized Twitter API to retrieve posts, and created a back end which we host with Linode. Then we formatted the text correctly from Twitter and pieced all of these things together with lots of tweaking to create the final website, which we built by hand with HTML5 (we wanted to learn how to use this language) and CSS, node.js, explorer, and JavaScript to make it work.
Challenges I ran into - Translating and making a Markov modeler in JavaScript, learning about HTML, Javascript, and back-end programming on the fly - this was out first hackathon, bugs, getting Twitter post collection to work properly.
Accomplishments that I'm proud of - Getting something created when new to so much of what we needed to built with - learning a ton!
What I learned - How to code in JavaScript and HTML5, how to create a basic backend, how to use node.js and how to host on a local server and on linode.
What's next for Twitter Oracle - Let Others Complete The Sentence - Enable the user to search any Twitter username and then have that person complete the sentenc
Log in or sign up for Devpost to join the conversation.