Text Mining OKCupid Data to Predict Users' Education Level

by Timothy Nguyen and Julian Marks

We applied text mining techniques to anonymized OKCupid Data in an attempt to predict users' education levels. This was done as a final project for PSTAT 131, UCSB's undergraduate Data Mining course. Our main question was whether or not the way people write to describe themselves carried any information regarding their level of education. Although our prediction results were far from glamorously accurate, we were able to conclude that users' writing did indeed carry information about their education level. For more details, please read the full report here.