John Erickson is a PhD student in Data Science at Worcester Polytechnic Institute. Before coming to WPI, John completed his undergraduate at Michigan State and his masters at Northwestern. His research interests include educational data mining and predictive analytics.
I got started off doing my undergrad at Michigan State and in economics and actuarial science. I did my masters at Northwestern where I shifted into the realm of predictive analytics. Predictive science combines statistics, mathematics, computer science and business to provide interdisciplinary smart decision making tools. Now, I am at WPI pursuing a PhD in data science.
My research now has pertained to utilizing natural language processing with math questions and open responses. I am analysing models to see how well our computers can understand how kids answer open questions. Our goal in this project is to take open responses and predict student grades. By developing this model, we will be able to give teachers who use ASSISTments a tool to increase grading efficiency. In the future, we are looking to provide responses to student answers in a fashion that resembles Google Smart Reply.
Part of this goes back to my previous research. At Northwestern, I used predictive analytics tools to study student performance in higher education. When I came to WPI, I took a course with Dr. Heffernan and was really inspired by his passion for K-12 research. I was interested in using the ASSISTments platform to ask important questions about how people learn.
I was specifically interested in personalizing education. I think it is important for education to go from being super standardized to more adapted to individual students. Being able to take my background in predictive science and data analytics applying to how students learn and helping teachers that idea really excited me. I think that people learn in different ways and at different rates. If we can analyze these different ways, we can find ways to help people learn even more.
For me, I think the big thing is to personalize education so that activities are adapted to students. I think applying data approaches to learning can have powerful impacts. Personalized learning features like the hints and feedbacks in ASSISTments allow students to approach a topic in their own way. Data analytics help to provide that personalization to students. Being able to understand student behavior and learning outcomes has the potential to transform education.