Overview

How we understand and communicate to those around us shapes and determines the person we are. In order to help people better understand the sentiments behind their words, we have incorporated the NLTK sentiment analysis tool to learn and interpret the positive, negative, and neutral meanings behind our words. In addition to d3js illustrations, integration in Moxtra allows for a commonplace setting for different users to benefit from the app.

The data is provided in three graphs: sentiment analysis of messages sent, sentiment analysis of messages received, and sentiment analysis over the course of using the website in hopes to help better become more aware of their words and surroundings, to improve the ways that people communicate, and to provide useful and meaningful data where necessary.

Challenges

The areas that we spent the most time on were actually in planning, incorporating Moxtra, and deploying to Azure. We actually encountered an interesting bug with the Moxtra API with user authentication after hours of slamming against it and had a great experience bringing it to the attention of one of the Moxtra engineers to find a workaround. Azure server setup with Flask was streamlined once talking with a helpful mentor - with most of the initial issues cause by hardware setup differences.

What's Next?

Mobile, more robust messaging, stronger analysis tools are all awesome features we would like to see! However, when we narrowed down the features to one item we would like to work on, we decided ultimately on the ability to retroactively review ALL messages in order to obtain an instant break down of communication over time and to observe any trends over time.

Share this project:
×

Updates

George He posted an update

Whats up next for Sentiment Chat? Expanding analysis tools to provide more concrete and solid results, retroactive analysis of all past messages in order to push usability - in addition to expanding to adopt a more robust messaging system. Incorporation of other important metrics such as time between messages into the learning model and dynamic analysis given context of messages will be sure to keep the balance of information both engaging and informative to our users!

Log in or sign up for Devpost to join the conversation.