ResearchSearch

• 

  Frontend

Inspiration

As someone interested in pursuing graduate school, looking into research areas and the professors who do that research is something that can be a somewhat daunting task. I created Research-Search to help people find research areas and professors with a smooth graphical interface.

What it does

Research-Search will scrape the inputted URL to a website listing university professors and find the bios of any listed faculty. From there, we utilize PCA and K-Means clustering to create a visual of the professors and how they relate to each other.

How we built it

The web app was created utilizing a dockerized FastAPI backend that uses BeautifulSoup to scrape an inputted university URL. The visualizations are done using d3.js with the data from the K-Means clustering to display the graphics on the frontend. The backend utilizes a Redis database for caching results.

Challenges we ran into

I had many issues trying to figure out how to display the graphical representation on the frontend. I eventually found something I was happy with after lots of trial and error.

Accomplishments that we're proud of

I was very proud to create a visually enticing frontend that allows the user to see how closely certain professors' research areas relate to each other.

What we learned

I learned about various unsupervised machine learning techniques (PCA and K-Means) and how to implement them within a scalable web system.

What's next for Research-Search

I would like to improve the visualization to include more than just the people listed on the website, but also people who have worked alongside these professors (e.g. common co-authors at other institutions). Additionally, I would like to add more detailed information about the clustering, perhaps using topic modeling techniques.

Built With

• docker
• fastapi
• python
• redis
• scikit-learn
• tailwind