There is a rising trend in using data to determine student performance and provide timely intervention for low-performing students/at-risk students. This project hopes to help highlight and contribute to these efforts in education innovation.
Having previously worked at a classroom learning/management platform edtech startup, I was inspired to do this project as a way to learn more about the process of building a data-driven student intervention system as well as to understand how data science can be used to transform education.
The dataset below was provided by Kaggle, and gives a snapshot of student engagement and student background as well as their corresponding final grades (classified by 3 categories: high-level grades, middle-levelgrades, & low-level grades.
This project aims to determine which factors are the greatest indicators in identifying a student as being at‐risk either behaviorally or academically and to make predictions, based on these factors, of which students belong to which grade class (high, middle or low).
Please view notebook for more details!