Student Professor Recommendation System

• 

  AWS SageMaker Screenshot of the Model running on a test data(A student's interests) and result is shown(Relevant Professors) Example #1

• 

  Test data Example #2

• 

  Test data Example #3

• 

  Test data Example #4

Inspiration & What it does

Introducing the Student-Professor Recommendation System, a cutting-edge platform designed to bridge the academic gap! In today's vast educational landscape, students often struggle to find the right mentors and research leads. We are firstly web scrapping all the Professor's data on their personal websites. This data contains information about past research works and current interests of all the Professors. We have then built a recommendation model using tfidf vectorization (And using a lot of data processing). Now if a student posts their interests (It can be very specific also), we recommend Professors who he/she can reach out to further contact them that we provide email, and office location.

How we built it

Diving deep into the project we have used AWS Comprehend to understand the main keyphrases from Professor's scrapped data. Then we used AWS Sagemaker to build the recommendation Model. Tried and tested with sample student test data. And we received relevant professor names and contact information as output for any particular interests string. We further created a web page, and hosted on AWS Amplify. We have the codebase on AWS CodeCommit and as soon as the student enters the details, it will trigger a AWS Lambda via AWS API gateway and store the form details in DynamoDB.

Challenges we are stuck at

The connection part of front end web page to the backend is partially unresolved at the moment(AWS API gateway). However the recommendation model is built completely in AWS SageMaker and gives accurate results.

Accomplishments that we're proud of

SageMaker model gives the expected results for any random student interest string if you enter you will find the relevant professors and their contact details on the SageMaker notebook output screen. So Student Recommendation system is built.

Web scraping is done to fetch all the Professor's data from their personal websites using George Mason University portal to get a directory of all the professors.

We have used AWS CodeCommit and AWS Amplify for the first time and we were able to successfully deploy the entire webpage on it and able to view via the following link. Webpage: https://main.dsp0p89pj860t.amplifyapp.com/

What we learned

We learned many AWS services, and which service should be used for what use case.

What's next for Student Professor Recommendation System

As mentioned in the challenges, we would like to continue building the front-end part of this project so that users can use it directly on the web page.

Built With

• amazon-dynamodb
• amazon-web-services
• amplify
• codecommit
• css
• html
• iam
• javascript
• lambda
• node.js
• python
• s3
• sagemaker