We were inspired to work on this project to serve to aid the miscommunication that's all too present in our world today. Leaders, professors, and authority figures of all shapes and sizes throughout the ages have had to determine their audience's sentiment for one reason or another while they garnered support or searched for a solution. Single individuals simply do not have the time or ability to interview each person of the audience to find out what they all thought and are usually left to make guesses of how they felt by cheers of support or boos of hate. Our goal was to give those over arching figures another tool to observe their audience to ultimately gain a better understanding of their ideas, feelings, and sentiment.

What it does

Sentithink attempts to solve this problem by using the concept of the Internet of Things to track, categorize, and visualize the audible sentiment of a geographically large group of people. Using multiple microphones dispersed throughout an area, Sentithink records 1 minute snippets of sound, transcribes it to text, parses out keywords, and tracks their frequencies over time. These frequencies, and their associated sentiments, are then visualized for the overseers to utilize to essentially see what's trending and how people are interpreting the event.

How we built it

Sentithink was built in three main parts consisting of microphone client side code, a Web API backend in Azure, and a front end visualization built in javascript.

Client Side: Microphones on machines running Python script to send snippets to Azure API endpoint Backend: Microsoft Azure Function API to record/produce results from speech to text Frontend: Javascript utilization of d3.js to show relative frequencies

Challenges we ran into

We weren't exactly sure how to use the microphones on the client side before hand but Google turned out to be a great resource. We did have some prior experience with Azure Cloud services but it seemed most of our trouble came from trying to visualize our data in javascript at the very end.

Accomplishments that we're proud of

 We were able to set up all Azure aspects of our program:
     SQL Database
     Azure Function Web App
     Front end visualization
     Proof of concept demo
     Lightweight clientside app

What we learned

We got a better understanding of Azure web services, javascript, and python utilization in a connect API driven environment

What's next for Sentithink

 Get actual IoT devices to scale the product and test our product on a large area

Built With

Share this project: