Study Buddy

Study Buddy

An all in one solution for schools and students to manage various academic and non-academic problems using machine learning algorithms and incentivizing students for growth by rewarding their efforts

The problem Study Buddy solves

Teachers are often faced with the problem of identifying in which subjects are students weak. Other than identifying the task of incentivizing students to improve also remains. Our solution not only solves the above mentioned problems, but also focuses on holistic development. It classifies students into two different sets; one being on a quadrant of Studies Well, Scores Well; Studies Well, Doesn't Score; Doesn't Study, Scores Well; Doesn't Study, Doesn't Score. On the basis of above quadrants we identify the strategies to apply on these students which include giving weekly tests on our platform, scheduling remedial lectures, identifying online courses etc. It also identifies which subjects a student is weak in which pairs well with above mentioned classification. It incentivizes students to perform well by giving them various opportunities like scholarships etc. Not only are we cheering students who perform well but also the students who show good growth. We perform typical student information system management like recording their attendance, various documents etc. We identify students that are exceptionally well in their fields and let them be part of various communities that suit their interests. Mental health of students is also taken into consideration, we ask them how they feel daily and then by using an NLP model identify the moods and feelings of the students. Taking into account student's future decisions we take their aptitude tests and find out how and where they can improve and also how to select their career path.

Challenges we ran into

One of our first challenge we ran into was actually finding a dataset that suit our needs, we eventually found one out on Kaggle. Identifying which subjects the student was weak at also became challenging as merely finding the minimum marks wouldn't have been enough. Instead we used an ML model first and then performed further analysis to find out the subject scored bad marks in. In our NLP model we had slight difficulty in finding out more emotions. Because our app consisted of so many API routes. Inegration of all was a real challenge for us.

Tracks Applied (1)

AI/ML

The part of identifying students that they belong to which criteria and identifying weak subjects falls into this domain...Read More

Discussion