Website
App

Project Title

HabitTracker

Description

Website- HabitTracker can record the number fo stress symptoms or "bad habits" you exhibit (for example, biting your nails, scratching your head until the blood oozes out- all symptoms of mental stress). HabitTracker uses real time hand detection directly from your browser. Underneath, it uses a neural network that provides bounding box predictions for the location of your hands in an image (using the Tensorflow object detection api). You can see the rendered predictions (a blue box surrounding your hand) and a variable confidence threshold for the machine learning model. When exhibiting any of these stress related symptoms. the "number of bad habits" increases and an audio plays, Michael Jordan encouraging the user to stop. We can submit different sessions and see the different sessions in your profile.

App- the app allows you to take control of correcting your bad habit. you can record your bad habit, the date at which is was performed, and the frequency of your actions. by doing this, one becomes mindful and self aware of themselves. most bad habits are a result of stress, which is why the app also allows you to record the stress levels of your day from a range of one to five inclusive. it even lets you add a picture. your picture can be anything. it could be something that calms you down, or it could even be you performing your bad habit so that you can see the frequency of your actions over long periods of time. the app is akin to basic journaling which is known for reducing stress and mental and emotional strain. the app shows you how many times you have consciously performed your bad habit whereas the website show you how many times you've done it (consciously and unconsciously). by pairing these two together, one can compare the frequency of their actions throughout the day and can accurately gauge the amount of consciousness they need to put into their day to correct their habit.

Technology Stack

Express JS & Node JS backend, html, css and vanilla js frontend and mongoDB database. We also used EJS templating and tensorflow object detection api. For the app, we used swift.