Our Story

The Instructional Equity Observing Tool is an online video/audio analysis tool that is geared towards assisting the teachers and faculty of educational institutions in analyzing and understanding how their interaction with students translates into real learning. Our platform is meant to replace the previous, manual method of analysis that many teachers/instructors perform to try and quantify different metrics about their teacher-student interaction. Instructors have expressed desire to view metrics such as the time the teacher talks during a lesson, what is the response time of students to those questions, and other data points such as the types of questions being asked (as categorized by Bloom’s Taxonomy). Quantifying these instructional variables helps these instructors more accurately understand the areas that they are strong in, and more importantly, the areas in which they can be more interactive with the students as to allow them to better absorb the lessons being taught.

With the help of our tool, we can allow teachers to quickly and efficiently gather this data about each of their lessons so that data driven changes in teaching techniques is possible, and moreover, so that teachers can identify potential vectors of ineffective instruction.

The process for using this application is for a user to login/sign-up for our site, then they will proceed to upload either an audio or video file to the designated location. Our tool will then take that video/audio file and execute a customized API call to AssemblyAI (https://www.assemblyai.com/) that transcribes this file into text. We then perform specialized data manipulation operations on the transcript to generate all the different metrics and display them in an easy-to-read format that the user can then scroll through and analyze the results. The user will also have the option to save this report that is generated as a pdf, which they or an administrator role will be able to access and view again at a later time.

Our application is hosted using Amazon Web Services (AWS) and utilizes many different functionalities that this service provides. AWS manages our authentication and authorization, user account management, and report storage functionalities. Our current system does not use its own machine learning model and instead offloads transcription to the AssemblyAI API, however this could be updated in the future with the addition of large datasets for training. A specifically trained machine learning model in this case could provide a more accurate categorization of questions and a more flexible tool that could eventually make predictions or suggestions to the user on the best ways to improve their teaching methods.

Meet the Team

Micah Collins

Data Visualization and Deployment

Rory McCrory

Frontend and Documentation

Sam Callan

AWS and Frontend

Yilika Loufoua

Frontend