Inspiration
We wanted to build something that was able to produce interesting visualizations that would essentially be a playground for people to try different inputs and see the results.
What it does
It takes in a text or audio input, finds words that are closely related semantically, and produces a word cloud to display the closest words.
How we built it
The frontend was built in react, using open source components for the word cloud generator and the audio recorder. The backend uses python's Natural Language Toolkit library and the wordnet lexical database, and looks through each word in the input to find words with similar definitions. The front and backend communicate with each other through jQuery ajax requests, and react components share state through redux.
Challenges we ran into
One of the biggest challenges was setting up the Ajax requests so that we could send data from the browser to the backend. Additionally, the algorithm for finding similar words was difficult to optimize in a way that wouldn't take too long for the user to see results. Unfortunately, near the end, we ran into an issue with accessing json values that we couldn't find a solution to, and so we needed to use a workaround that wasn't great.
Log in or sign up for Devpost to join the conversation.