As graduate students with interest in research, we feel there is a need for a system that matches professors to students to work in the same projects. To our knowledge, currently there is not such system that does this kind of matching based on research interest or statement of purposes. We pulled our inspiration from dating apps and built a text mining based application that matches students with professors.
What it does
Our pipeline takes in research interest from students in plain text and recommends professors based on their profiles, rating and project description. We are using sematic search to compute the ranking. We are also adding weights to ratings and their funded projects.
How I built it
We crawled the data from UMBC IS department. We took profiles of Full-Time professors, descriptions of currently funded projects, linked them with the PIs and Co-PIs and took ratings from RateMyProfessor. We built our UI R. We used LSA for semantic search using R and Python.
Challenges I ran into
- App development and integration
- Web crawling - web pages did not follow any given structure
- Integrating R and python
Accomplishments that I'm proud of
We validated with current students and professors and pretty accurate!!!
What I learned
We put the concepts of data science such as dimensionality reduction, semantic search into practical use
What's next for MeetYourProf - Find your research match!
We plan to crawl websites of multiple universities, so the students can find professors across the country. We also want to do the reverse, where professors can find students.