Inside Out

Abstract

In the modern world, depression has become a major problem for society, and it is increasing day by day. Depression is resulting in a lot of deaths and has become a hurdle for people who want a normal happy life. People usually share their feelings and emotions through the social media platform. We have developed a desktop application that can help educational institutions or corporations to identify the presence of probable depression levels by analyzing their tweets or extracting the emotions from the facial expressions of the students or employees, respectively. We have assigned different percentages to the different probable depression levels. The tweets of the user are classified into positive and negative tweets based on different metrics. The admin of the system can scrape the tweets from the Twitter account of the user and apply sentimental analysis for the prediction of the probable depression level. The admin can also predict the probable depression level from the facial expressions of the user and compare the results with each other. The manager can record the live stream of the user and can generate monthly or yearly reports of the people to statistically analyze the results. We have compared the results generated from the CNN model with the logistic regression model. This can help the organizations to predict the probable depression level beforehand and take necessary precautions so that the person develops a better personality. This project is very beneficial for society because nowadays most of people are depressed due to various reasons and their posts on social media directly affect society.