View Project
For my DS 110 final project, I used various Python libraries (pandas, scikit-learn, nltk) to analyze a dataset of 20,000 RateMyProfessors reviews for characteristics that most impacted professor ratings. I trained random forest classifier and linear regression models to predict each professor’s average rating. I also experimented with Word2Vec embeddings to compute attractiveness ratings for professors based on students’ descriptions of them, and conducted t-tests to determine statistical significance of correlations (yes, more attractive professors had higher ratings).