-
Clustering Visualisation(a)
-
Clustering Visualisation(b)
-
Clustering Visualisation(c)
-
Compare ratings
-
Search by location
Inspiration
Choosing a coaching institute is a big decision that - if made hastily - can lead to crashing dreams. We help you match with the best institute out there for you.
What it does
We help a student match to an institute to help them pursue their choice of career path. With the huge options available out there, it's easy to get looped into a not-so-good institute based on word of mouth.
We used User-Item and User-User based collaborative filtering algorithms to give recommendations to the users.
As the number of pairs between user-user and item-item can be very large so it is necessary to break down, the large sets, into small clusters, which can then be pairwise. We have displayed the visualizations graphically for k-means and spectral clustering.
Compare Feedback given by users to different institutions, can be compared category wise.
How I built it
We used collaborative filtering, machine-learning, data analytics, to recommend the institutes to the user, it is a web based portal built using django for the backend, and jquery, javascript, html5, highcharts, bootstrap, css3 for front-end For visualisations, we have used matplotlib and scikit library.
Challenges I ran into
Comparing different clustering, reducing high dimensional data to two-dimensional data for visualisations. Personalising the recommender system. Dummy data generation for analysis.
Accomplishments that I'm proud of
Personal recommendation system, easy to use portal for layman, with pretty user interface and experience.
What I learned
Machine learning implementation, recommender system personalisation, Highcharts API, google maps api,
What's next for courseTube
Improve recommender system, generating some data Add more features, like enroll students
Log in or sign up for Devpost to join the conversation.