Livio – Your Smart Flatmate Finder

🚀 Livio is a multilingual roommate matching app that uses real-world data and cosine similarity to recommend the most compatible flatmates — based on lifestyle, habits, and personal preferences. Built entirely in Python with Streamlit, it’s designed to be explainable, beautiful, and exportable.

🔗 Live app → livio-roommate-matching.streamlit.app
📦 GitHub repo → github.com/Isaac5Tayfur/livio-roommate-matching

---

🔧 Key Features

• 🎯 Personalized matching using Cosine Similarity on normalized lifestyle traits

• 🌍 Bilingual UI (English and Spanish)

• 📊 Interactive charts and profile comparisons

• 🧬 Trait explanations based on shared lifestyle preferences

• 🧼 Filters for non-smokers, healthy eaters, and pet allergies

• 📥 Export results as CSV and high-res PNG (chart + table)

• 💡 Dynamic theming with responsive layout and dark/light mode support

---

🧠 How It Works

Livio loads a dataset of 15,000+ tenant profiles, each with traits like:

• Sleep schedule, work shift, energy rhythm

• Budget, social level, cleanliness rating

• Preferences like reading, pets, loud music, cooking, diet, etc.

These profiles are encoded and normalized with MinMaxScaler, then compared using Cosine Similarity.

You select up to 3 seed tenants, and Livio recommends the top N most similar profiles — based on shared normalized traits.

---

🖥 App Interface

Below are examples of the Livio interface in both English and Spanish versions:

🇬🇧 English UI

🇪🇸 Spanish UI

---

📊 Outputs & Exports

After running a match, Livio displays:

• A bar chart of compatibility scores

• A comparison table of key lifestyle traits

• An explanation block with top shared habits (e.g., 🛏️ Early risers, 🧼 Clean, 🚭 Non-smokers)

You can also download the results:

📋 Table (PNG)

📊 Chart (PNG)

---

🗂 Project Structure

Livio/
├── app.py               # Streamlit frontend
├── logic.py             # Matching algorithm & data preprocessing
├── ui_helpers.py        # Chart, table, and trait explanation functions
├── tenants_dataset.csv  # Full dataset of profiles
├── metadata.xlsx        # Schema of all traits and value options
├── Media/               # Visual assets (banner, exports)
├── requirements.txt     # Python dependencies
└── README.md            # GitHub documentation

---

🚀 Try It Yourself (Locally)

git clone https://github.com/Isaac5Tayfur/livio-roommate-matching.git
cd livio-roommate-matching
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt
streamlit run app.py

🟢 Python ≥ 3.9 required. No database needed — all data is in CSV format.

---

☁️ Deployment

Livio is deployed live using Streamlit Cloud.
No signup is needed — just open and try:
🌐 livio-roommate-matching.streamlit.app

---

🧭 Future Enhancements

• User authentication and history

• Weighted trait matching

• Database-backed profiles

• Additional languages (French, German...)

• Mobile optimization and saved preferences

---

👤 Author

Tayfur Akkaya Clavijo
🔗 LinkedIn | 🧑‍💻 GitHub | ✍️ Blog

---

📜 License

MIT License. Free to use, adapt, and contribute.