Choosing professors is difficult, especially among the equally great. Input the name of a memorable professor and match him/her with the professor most similar in personality.
How It Works
Our data model uses text classification on Polyratings reviews. It produces similarity scores between professors from the adjectives that students use to describe them.
Challenges We Faced
We first used GCP's Sentiment Analysis but found that it had trouble managing the size of our dataset. It also didn't provide a suitable metric for professor comparability. Instead, we aggregated the counts of most common adjectives and used a weighted model to estimate differences in personality.