🥗 NutriPulse

Your smart, real-time nutrition & fitness tracker

---

✨ What is NutriPulse?

NutriPulse is a full-featured, beautifully designed nutrition tracker built with Streamlit and powered by two nutrition APIs. Type natural phrases like "2 cups white rice with carrots" and instantly get accurate macro breakdowns, calorie tracking, exercise logging, and personalized insights — all in one clean interface.

Built as a course project, NutriPulse demonstrates real-world API integration, data visualization, and modern UI design patterns within a Python web app.

---

🖼️ App Preview

Today's Log	Insights	Profile & Goals
Calorie goal + macro bars	Nutrition Grade A–F	TDEE calculator
Water intake tracker	Nutrient radar chart	Custom macro targets
NLP food search	Meal composition donut	Recent foods history
Exercise tracker	Smart suggestions	Calorie trend chart

---

🚀 Features

🍽️ Tab 1 — Today

Feature	Description
Calorie Goal	Set a daily target; colour-coded progress bar goes green → yellow → red
Macro Mini-Bars	Live protein / carbs / fat / fiber progress vs your custom goals
Water Tracker	Log by glass, 500 ml, or 1 L; adjustable daily goal
NLP Food Search	Type "1.5 tbsp peanut butter" or "bowl of oatmeal with banana" — it just works
Dual API	CalorieNinjas NLP is primary; USDA FoodData Central is automatic fallback
Inline Macros	Full nutrition breakdown (kcal, carbs, protein, fat, fiber, sugar, sodium) visible before adding to log
Meal Categories	Breakfast · Lunch · Dinner · Snack — colour-coded groupings
Food Log	Grouped by meal with per-meal calorie totals + stacked macro bar
Export CSV	Download today's full log as a spreadsheet
Exercise Tracker	10 activity types × MET-based calorie burn; net calories auto-updated

📊 Tab 2 — Insights

Feature	Description
Nutrition Grade	A–F letter grade based on calorie adherence, protein, fiber & macro balance
Smart Suggestions	Rule-based tips for low protein, low fiber, dehydration, over-goal calories
Radar Chart	Spider chart — today's intake vs your 5 custom nutrient targets
Meal Composition	Interactive donut chart — % of daily calories per meal type
Food Comparison	Search two foods side-by-side; grouped bar chart + diff table

👤 Tab 3 — Profile & Goals

Feature	Description
User Profile	Name, age, gender, weight, height, activity level
TDEE Calculator	Mifflin-St Jeor BMR × activity multiplier; one-click "Apply as goal"
Custom Targets	Set personal goals for protein, carbs, fat, and fiber
Recent Foods	Your last 10 searches for quick re-use

📅 Tab 4 — History

Feature	Description
Daily Snapshots	Save today's totals with one click
History Table	All saved days in a clean data table
Export History	Download full history CSV
Calorie Trend	Line chart showing calorie intake over saved days

---

🛠️ Tech Stack

Frontend     Streamlit 1.48+   Python web framework with reactive state
Charts       Plotly             Bar, Radar, Pie, Scatter charts
Data         Pandas             DataFrame manipulation & CSV export
Styling      Custom CSS         Inter font, green gradient theme, card layout
API (Primary) CalorieNinjas     Natural language nutrition parsing
API (Fallback) USDA FoodData    380,000+ food database, relevance scoring
Math         Mifflin-St Jeor   BMR & TDEE formula for calorie recommendations

---

⚡ Quick Start

1. Clone the repo

git clone https://github.com/VanshikaGurbani/NutriPulse.git
cd NutriPulse

2. Install dependencies

pip install -r requirements.txt

3. Set up API keys

Create a .env file in the project root:

USDA_API_KEY=your_usda_key_here
CALORIENINJAS_KEY=your_calorieninjas_key_here

API	Free Tier	Get Key
USDA FoodData Central	3,500 calls/day	api.data.gov
CalorieNinjas	10,000 calls/month	calorieninjas.com/api

4. Run the app

streamlit run nutrition_tracker.py

Open http://localhost:8501 🎉

---

☁️ Deploying to Streamlit Cloud

1. Push to GitHub
2. Go to share.streamlit.io → New app → select this repo
3. In App settings → Secrets, add:

USDA_API_KEY = "your_key"
CALORIENINJAS_KEY = "your_key"

4. Deploy — live in ~60 seconds

---

📁 Project Structure

NutriPulse/
├── nutrition_tracker.py   # Main app (backend + UI, ~750 lines)
├── requirements.txt       # Direct dependencies only
├── .streamlit/
│   └── config.toml        # Green theme configuration
├── .env                   # API keys (gitignored)
└── README.md

---

🔬 How the Food Search Works

User input: "2 cups white rice with carrots and peas"
         │
         ▼
  CalorieNinjas NLP API  ──► Returns each ingredient separately
  (primary)                   with quantities already parsed
         │
    No result?
         │
         ▼
  USDA FoodData Central  ──► Relevance scoring algorithm
  (fallback)                  penalises wrong-category matches
                              e.g. "Bread, oatmeal" ≠ "oatmeal"

The custom USDA relevance scorer uses:

• First-token match bonus for foods that start with the query word
• Wrong-category penalty (-10) for USDA "Primary, query-word" naming patterns
• Cooked/raw context bonus/penalty based on food type
• Calorie sanity check for unusually high per-100g values

---

🎨 Design System

• Font: Inter (Google Fonts)
• Primary colour: #3cb371 (medium sea green)
• Gradients: #2d6a4f → #40916c → #52b788
• Cards: st.container(border=True) + custom CSS targeting [data-testid="stVerticalBlockBorderWrapper"]
• Macro colours: Calories=orange · Carbs=amber · Protein=blue · Fat=purple · Fiber=green

---

📜 License

MIT — free to use, modify, and distribute.

---

Built with 💚 using Streamlit · Data from USDA FoodData Central & CalorieNinjas

🚀 Try it live →