In short:
I scraped data from Rate My Professors and analyzed it to find the factors that influence a professor's rating.
Introduction
RateMyProfessors.com is a website that has student ratings and reviews for professors and the various courses they teach. It is widely used by students when selecting courses with multiple sections and different professors to decide which professor to go with or to find out what to expect when they take a class with a particular professor.
I investigated the different factors (e.g. course clarity, difficulty, textbook usage) that influence a student’s satisfaction with a course and, thereby, the ‘rating’ they give the professor. I also wanted to investigate the impact of online learning on these same factors.
For my analysis, I used all the student ratings and reviews for the university of Ottawa courses and professors available on RateMyProfessors.com and analyze them using various methodologies, including hypothesis testing, correlation coefficients, predictions and features importance.
Findings
From my initial tests and analysis results, I concluded that the most important features contributing to a professor rating are:
- Clarity of the professor and course
- The grade a student gets in the class
- The student's interest in the class
- How easy the class was
I also found that students found online courses to be less clear, and gave them lower ratings, but performed significantly better.
Ethical concerns
Finally, the main ethical concern I found while doing this project was protecting the professors' identity in our datasets. I made sure to leave out any names and replace them with IDs. This prevents from associating reviews with a professor and protects their identity
Another ethical concern is the issue of no rate limiter on RMP API. API rate limiting is a common practice that is used to ensure that an API is not overwhelmed with too many requests, which can lead to performance issues and potentially even downtime. Currently, RMP API has no rate limiter which means anyone with bad intentions could abuse their servers.