🏡 Property Recommendation System

This project is a Python-based property recommendation tool that uses feature similarity to identify the top k comparable properties to a given subject property. It reads property data from a JSON file and outputs the most similar properties using a distance-based algorithm.

---

📂 Project Structure

• training_dataset.json: Input file containing subject property, candidate properties, and known comps.
• output.csv: Output file with the top k most similar properties.
• software_recommendation_notebook.ipynb: Jupyter notebook containing the full workflow from preprocessing to distance computation and output.

---

⚙️ How It Works

The algorithm:

1. Parses and cleans input data (e.g., GLA, Year Built, Structure Type).
2. Encodes categorical data (one-hot encoding for structure type).
3. Scales numerical values using MinMaxScaler.
4. Calculates Euclidean distance between the subject and all candidate properties using feature weighting.
5. Returns the top k most similar properties and saves them to output.csv.

---

📊 Features Used for Comparison

• Gross Living Area (GLA)
• Total Room Count
• Property Age (calculated from year built or age value)
• Structure Type (one-hot encoded)

---

🧮 Distance Weights

Feature	Weight
GLA	1.0
Rooms	1.0
Age	1.0
Structure Type	0.2 (each encoded column)

---

📥 Input Format (training_dataset.json)

Required Keys

{
  "subject": {...},
  "properties": [...],
  "comps": [...]
}

• subject: The target property you're trying to compare against.
• properties: A list of all potential comparable properties in the area.
• comps: A list of previously chosen comparable properties. These will also be evaluated alongside properties but flagged in the results.

---

▶️ Usage

Open the notebook using Jupyter:

jupyter notebook Software_recommendation_notebook.ipynb

Run each cell sequentially to perform preprocessing, compute similarity, and view/save results.

You can modify the number of recommendations by changing the value of k in the notebook.

---

💾 Output

The output is saved to output.csv, containing:

• Rank
• ID
• Distance from subject property
• Address
• Is Comp (True/False)
• Full property details (including GLA and bedroom count)

---

📦 Dependencies

Install required Python packages with:

pip install pandas
pip install numpy
pip install scikit-learn
pip install scipy

Python 3.7+ is recommended.

---

📘 License

MIT License