Predicting K-12 Graduation Using Machine Learning
Program: Data Science Master's
Host Company: Public K-12 school district in Washington State
Location: Greater to Seattle Area, Washington (hybrid)
Student: Nicholas Teal
The goal of this project was to try and predict which students will or will not graduate on time purely based on data from their 9th-grade year. It is pretty obvious that if a student fails a bunch of classes they are less likely to graduate, but are there any other predictors that have a strong relationship with graduation that are not obvious? By using a penalized regression we can create a model that is interpretable with fewer predictors while still being accurate. This model can be used for future students to find students who are off track to graduate that might have flown under the radar and not received the interventions they need.