Our Idea based on Depression which is a mood disorder that can affect a person's daily life. It may be described as feelings of sadness, loss, or anger. As of 2017, 300 million people around the world have depression, according to the World Health Organization. It's important to take action against depression - it doesn't just go away on its own.Being depressed can make you feel helpless. You're not.

What it does

It's just a small initiative against depression. Not all people but some people put Depression quotes in stories of social media or in tweets. So, We trained our machine learning model like this you just input that quote in our app and it will tell you the person is sad or depressed so you can talk to him/her about problem. We also integrate our chatbot using Google Dialogflow.

How we built it

We take our first step for the machine learning model so first we import the necessary libraries used for this project and then we load the dataset. After this we separate message and label column and done data preprocessing and data cleaning of message column and then vectorized the text using CountVectorizer and split the data into train and test with 80% and 20% size respectively. After this we train the model using Multinomial Naive Baiyes Algorithm and achieve the 96% accuracy. Finally dumb the our model in pickel file and wrote the backend in flask and make it visual in frontend using HTML, CSS and Javascript. We train our chatbot using dialogflow and attached to our Web Application. So if the person is feeling depressed he/she can chat with out bot.

Challenges we ran into

We ran some challenges into deployment because it is our first time we are using Heroku for deploying our app so we search how to deploy it and failed 3 times because of some heroku error then after approx 3 hours of grilling we've finally done it.

Accomplishments that we're proud of

We're proud of that we made a full stack Web Application with integration of Machine Learning and we made something about that topic which is a big issue but no one talks about it.

What we learned

We learned how can we integrate Machine Learning Model in our Backend and use it in our website.

What's next for HaSa (Happy-Sad)

We are thinking to train our chatbot more so it will help people more and thinking to create an API so if anyone wants to use it in mobile development so he/she can use it.

Built With

Share this project: