Inspiration

The inspiration was to build a safe platform for all students, sometimes a student is hesitant to reach out for help in person or they cannot due to the size of their class. With We connects all students can get the help they need.

What it does

In Vconnect students are connected with other students in their classes, it is a safe platform where they can host virtual meetings and organize group studies.

First, they create a profile where they can share information on what classes they are taking, their previous classes and also their skill sets. They also get to choose their best skills and highlight them on the top of their profile.

Based on the classes that each student is taking, our machine learning model connects those classes with the skills of individual students. So in this way, they can get recommendations on who they can request for help on a virtual meeting. We have embedded Zoom and Gmeet for virtual meeting Platforms. As soon as a virtual meeting is requested, the selected student receives an email informing them about this request and a link to the zoom meeting. This was done using the Twilio API.

After a meeting is scheduled, the student profile gets updated. It shows all upcoming meetings and details of each meeting on clicking them.

I have paid particular attention to User experience, to make sure that no-one gets confused. The begin meeting button stays disabled until 30 minutes before the meeting starts. Once a meeting is hosted other students can also request access to it. All meetings are sorted based on their time of hosting.

Also 30 minutes Before the meeting students will get another email and text reminder. It was also done using Twilio. Lastly, wanted to ensure that all students are in a good mental space. Wanted to make this platfrom a source of relieving their stress and create a safe student community. so I have also integrated an option to hold non-academic meetings which students can join for some fun. These meetings are specially designed to address each students interests like LGBTQ+, and Happy Hour where students can have lunch together.

Using Twilio API and zoom API together we have made an easy user interface as it automatically creates a Gmeet or Zoom link and sends it through an email. We used over 200 articles from Studentsaffairs.duke.edu, for match and recommendations. We have also connected MongoDB as a database to save all the information and integrated Twilio, sendgird, Esri, meetup, zoom, and Calendly API.

How I built it

In short: 1) We build our front-end from scratch and it's beautiful and polished using bootstrap, JavaScript and Flask. 2) Used Twilio API to send email to the selected student for asking request. Automatically sending reminder 30 minutes before meeting is about to begin. 3) Blackened Integration of Many API's Twilio, sendgird, Esri, meetup, zoom, and Calendly API.

Details: Front-End and Back-end: The entire front-end was built from scratch using JavaScript, html and css and we integrated it with our node.js back-end. We used twilio api to automatically trigger email notification to the attendees of the meeting and also send an essential remainder a few minutes before a meeting in their schedule starts. I used zoom and google-talk api to enable users to create meetings and redirect to these platforms from the schedule. I used calendly api to create user meeting schedules In our back-end we implemented mongodb atlas to handle our database operations like fetching user profile, skills, interests, location, email id.

Meetup recommendation system: To build the model we first webscraped April month’s information about virtual meetups using the Meetup.com API. For each group, queried the organizer ID, category tags, location, the topics covered by it, and its member count then built a data analysis pipeline using Python on AWS EC2. Then chose Factorization Machines model which is a supervised machine learning technique and works great with large sparse datasets. It uses both the feature interactions and the group features. To build the system, we divided the data into cross-validation (CV) and holdout set, trained the model using the CV set. A content based RS was used since we had to use the categorical features of each group.

Challenges I ran into

Integrating front end and Machine Learning model with the back-end.

Accomplishments that I'm proud of

To be able to accomplish a complete Machine Learning based web platform and able to integrate all the libraries

What I learned

Using Twilio API Dealing with MongoDB combining the recommendation model with the front-end. Product planning Time management

What's next for We Meet

To host the platform on the server. To add another feature of having in person virtual meetings with professor.

Share this project:

Updates