Hack Health ++

Inspiration

Depression is a serious mental illness that requires understanding and care, otherwise it may be lead to life-threatening consequences such as self-harm and suicide. Depression cures are widely available, and with professional help, depression can be treated. If depression tendencies in a person could be identified at an early stage, the necessary help could be given so that the situation does not become worse. Language is a major component of mental health assessment and treatment. Today, many people express their thoughts and emotions on social media, depression being one of them. Thus, social media language can serve as a useful lens for mental health analysis.

What it does

In this work, we seek to combine the power of big data and deep NLP models to predict whether the tweets express a depression tendency or not. We consider a dataset scrapped from twitter and try to explore if the Tweets can be classified as depressed or not. This works present a comprehensive study of different state of the art models to classify a tweet and to evaluate the shortcomings of different model in different situations. We detect depression tendencies in a twitter user and let them monitor their mood changes over a time thorugh an App. Also, professional help options provided in case of severe depression.

How I built it

We scraped the data for training our model from Twitter and existing datasets. We trained NLP models using Keras and used Flask for creating APIs that would predict depression in Tweets in real-time using the trained models. We also created an Android app for helping the users to monitor their mood changes over a period of time.

Challenges I ran into

One major challenge was obtaining the data. We had to search for datasets where users had already annotated depressed Tweets. Another major challenge was to integrate the our Android app with the backend which we solved by hosting the backend on cloud (AWS EC2).

What I learned

I learned how important is to set clear objectives when working on a tight schedule with a team you are not familiar with, the challenges of development increase in this scenario. We got to explore various ideas related to mental health which is the theme of our hack