- Extracted around 10k tweets using Tweepy and annotated them manually as depressive/non-depressive for applying supervised learning algorithms. - Applied various supervised learning algorithms like Naive Bayes, SVM, KNN and Logistic Regression. Achieved highest accuracy of 78% using Logistic Regression and highest precision of 0.8 using SVM. - Analyzed trends in twitter users by extracting their previous tweets and predicted the onset of depression in user. This can prove to be very helpful in early detection of depression trends in users.