We were planning on how to go forward with our project from last Pennapps link (a web app that donates leftover food from restaurants to homeless shelters) and we wanted to know how organizations with similar aims have gone about it in the the past. Moreover, we wanted to know how efficient they were, and how their actions were perceived. What we wanted was an opinion. Then it hit us that we're not the only ones looking for an opinion. So many companies spend billions of dollars on consumer research to find out whether the public loves or hates their products. We forget that Twitter is filled to the brim with the very data we're looking for: opinions on literally everything.
What it does
Emogram is a web app that performs sentiment analysis on recent tweets using Natural Language Processing (NLP) and aggregates this data to put on a map. The map shows the population density of each emotion connected to the topic at hand (be it a company's product or any other issue).
How we built it
We took tweets from twitter using the Twitter Search API and sorted them by location. We used the IBM Watson Alchemy API to analyze the tweets for their emotion. We used d3.js for the data visualization aspect, Linode to host our whole web app.