Feelee: Personalized Mental Health Support for Young People

General Description

Problem

Young people are increasingly experiencing stress-related complaints, but traditional mental health care often fails to attract this demographic. There is a significant need for accessible and personalized mental health support. Feelee aims to fill this gap by offering adaptive advice and interventions through smart-cueing. To date, over 4,000 young people have downloaded and used the app.

Goals

• Pattern Recognition: Conduct exploratory data analysis (EDA) and statistical data analysis (SDA) to find patterns among Feelee users.
• User Profiling: Develop a machine learning model to create user profiles based on the data.

Plans and Methodologies

Data Preparation

1. Clean the Dataset: Ensure the dataset is organized and ready for analysis.

Data Exploration and Analysis

1. EDA + SDA: Summarize statistics, create visualizations, and perform correlation tests to understand the dataset better.

Machine Learning Models

Algorithm: k-means clustering

1. Inputs:
   • User Information: Demographics such as age and gender.
   • Emotional States: Responses to questions about their emotional state.
   • Physical Activity: Data on physical activity, such as step counts.
2. Output:
   • User Profiles: Classify users into one of 6-12 clusters based on the inputs.
3. Result Analysis

Requirements

• Data Reduction: Focus on user information and the last 7 days of emotional states and physical activity.
• Minimum Data: At least 7 days of information is required.

Other Methods

Additional methodologies may be employed to enhance the analysis and model development as needed.