Student Mental Stress Assessment

Predict stress levels and get tailored coping suggestions—powered by machine learning.
By Leela Dammalapati, with contributions from Utkarsh Tripathi.

---

Try it online

No setup required—use the live app here:
→ mental-stress-assessment on Render

---

What it does

This web app helps students understand their current stress level and suggests evidence-based coping strategies. You sign up, complete a short assessment (demographics, habits, academic load, support, etc.), and receive:

• A stress classification (Low / Medium / High) with confidence scores
• Top 5 coping mechanisms recommended for people with similar profiles

Everything is stored per user, so you can track how your stress and recommendations change over time.

---

What’s included

Area	Description
Auth & profiles	Sign up, log in, and manage multiple profiles (e.g. for different family members).
Predictions	Random Forest model predicts stress level from ~20 input features.
Recommendations	k-NN model suggests coping mechanisms from similar users (k=50).
History	View past assessments and trends.
UI	Responsive, gradient-style interface that works on desktop and mobile.

---

Stack

• Server: Flask 3.0, SQLAlchemy, Flask-Login; production DB: PostgreSQL
• ML: scikit-learn (Random Forest + k-NN), pandas, numpy, joblib
• Client: HTML5, CSS3, JavaScript; no heavy frameworks

---

Run it locally

You need: Python 3.11+ and pip.

# Get the repo
git clone https://github.com/utkarsh9630/mental-stress-assessment.git
cd mental-stress-assessment

# Isolate dependencies
python -m venv venv
source venv/bin/activate   # on Windows: venv\Scripts\activate

# Install and configure
pip install -r requirements.txt
cp .env.example .env      # then edit .env as needed

# Start the app
python app.py

Then open http://localhost:5000 in your browser.

---

Repo layout

mental-stress-assessment/
├── app.py              # Flask app and routes
├── models.py           # SQLAlchemy models
├── forms.py            # WTForms
├── templates/          # HTML (base, auth, dashboard, assess)
├── models/             # Trained artifacts (rf_model, knn_model, scaler, *.json)
├── data/               # train_recs.csv etc.
├── requirements.txt
├── Procfile            # for Render
└── README.md

---

Hosting

The demo runs on Render with a PostgreSQL database. To deploy your own instance, use the button below (or follow Render’s docs for Flask + Postgres).

---

Roadmap

• Email alerts when stress is high
• Export assessment history as PDF
• Mobile app
• Wearable / external data integration
• More languages

---

People

• Leela Dammalapati — original author
• Utkarsh Tripathi — contributor

We’re open to pull requests and ideas. See the repo:
github.com/utkarsh9630/mental-stress-assessment

Contact: Leela Dammalapati (original author); Utkarsh Tripathi — tripathiutkarsh46@gmail.com

---

License: MIT — see the LICENSE file.