House Price Predictor

Quick Links

• Run locally
• Architecture
• API Examples
• Data & Model Notes
• Next Steps

---

Overview

This repository implements a small FastAPI web application that serves a single-page client and a /predict endpoint backed by a serialized scikit-learn pipeline. The app demonstrates a complete prediction flow: user inputs → model prediction → percentile against dataset.

---

Architecture

flowchart LR
  A[User Browser] -->|GET /| B(FastAPI)
  B --> C[templates/index.html]
  B -->|POST /predict| D{Prediction API}
  D --> E[Load pipeline: joblib]
  D --> F[Compute percentile using Housing.csv]
  E --> G[scikit-learn pipeline]
  style A fill:#f9f,stroke:#333,stroke-width:1px
  style B fill:#cff,stroke:#333,stroke-width:1px
  style D fill:#cfc,stroke:#333,stroke-width:1px

Loading

This simple diagram shows the request flow: the browser serves a static client, which posts JSON to /predict. The API loads a serialized pipeline and the CSV dataset to compute a percentile for context.

---

API Examples

Use the API directly from curl or any HTTP client.

curl example:

curl -s -X POST http://127.0.0.1:8000/predict \
  -H "Content-Type: application/json" \
  -d '{"area":2300,"bedrooms":3,"bathrooms":2,"stories":1,"mainroad":1,"guestroom":0,"basement":0,"hotwaterheating":0,"airconditioning":1,"parking":2,"prefarea":1,"furnishingstatus":0}'

Python example using requests:

import requests

payload = {
  "area":2300,
  "bedrooms":3,
  "bathrooms":2,
  "stories":1,
  "mainroad":1,
  "guestroom":0,
  "basement":0,
  "hotwaterheating":0,
  "airconditioning":1,
  "parking":2,
  "prefarea":1,
  "furnishingstatus":0
}
resp = requests.post('http://127.0.0.1:8000/predict', json=payload)
print(resp.json())

---

Data & Model Notes

• Housing.csv must contain the price column used to compute percentiles.
• main.py expects house_price_pipeline.pkl at the repo root. The object should support .predict(X).
• The Pydantic House schema defines the input fields and types.

Model training checklist (recommended):

1. Split dataset (train / val / test).
2. Build a scikit-learn pipeline with necessary preprocessing (scaling, one-hot or ordinal encoding for categorical fields).
3. Train, evaluate (RMSE, R²), and persist the pipeline with joblib.dump(pipeline, 'house_price_pipeline.pkl').

---

Run locally

1. Create and activate a venv, then install deps:

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Ensure house_price_pipeline.pkl is present.

3. Start the app:

uvicorn main:app --reload --port 8000

Open http://127.0.0.1:8000/ to use the UI.

---