BodySync+  :muscle: :heartbeat:

Personal Fitness & Preventive‑Health Suite powered by Machine Learning

“Tune your body, sync your life.”

---

🔑 What Is BodySync+?

BodySync+ is a Flask web application that blends data‑driven fitness with preventive health screening.
From a single dashboard a user can:

1. Estimate calories burned for any workout session.
2. Receive a personalised workout recommendation based on lifestyle.
3. Screen for five chronic diseases—Diabetes, Hypertension, Heart‑Attack, Stroke and COPD—using calibrated risk probabilities.

All models are trained on open datasets (Kaggle & NHANES), persisted via Pickle, served in real time, and wrapped in a Bootstrap UI for a smooth green aesthetic.

---

🖼️ Live Screenshot

---

⚙️ Tech Stack

Layer	Libraries / Tools
Back‑end	Flask 2.2 · Jinja2 · Gunicorn
ML	scikit‑learn (RandomForestRegressor / Classifier, MultiOutputClassifier, Isotonic Calibration)
Data	Kaggle Calories & Exercise · NHANES 2015‑16
Front‑end	Bootstrap 5.3 · Vanilla JS
Ops	Python 3.11 · virtualenv · (optional) Docker & GitHub Actions

---

📊 ML Models at a Glance

Page	Dataset & Size	Features	Algorithm & Tricks	Key Metric
Calorie Burn	exercise.csv + calories.csv (~1 k rows)	Age, Weight, Height, Heart‑Rate, Duration	RandomForestRegressor, 100 trees	MAE ≈ 12 kcal
Workout Recommender	Synthetic lifestyle dataset (2 k rows)	Age, BMI, Exercise Days/Wk, Diet Score, Sleep, Smoker, Alcohol	RandomForestClassifier with class_weight='balanced_subsample'	F1‑weighted 0.94
Health‑Risk Screener	NHANES (≈ 9 k participants)	12 vitals & lifestyle fields	Multi‑output RandomForest, isotonic probability calibration, custom thresholds	AUC 0.94 (Diabetes); recall 0.45‑0.80 across labels

Overall Health Score

We combine weighted disease probabilities into a 0‑100 Health Score (higher = healthier) so users instantly grasp their status.

---

🚀 Quick Start

# 1. Clone
git clone https://github.com/<yourname>/bodysync-plus.git
cd bodysync-plus

# 2. Environment
python -m venv venv && source venv/bin/activate   # Windows: venv\Scripts\activate
pip install -r requirements.txt

# 3. Get pre‑trained models (small zip, ~10 MB)
wget https://github.com/<yourname>/bodysync-plus/releases/download/v1.0/models.zip
unzip models.zip -d models/

# 4. Run
python app.py
# open http://127.0.0.1:5000