Housing prices prediction API

This API provides a prediction model, which estimates the price of a property given neighborhood such as number of households, population, proximity to the ocean etc..

Please read below to see how to run the application either using Docker (production) or Flask (development).

TL;DR

Run the application using docker-compose up and see the application on http://localhost:5000/.

The price prediction API endpoint is available at the http://localhost:5000/v1/predict-price URL. Note that you have to use POST request with the valid JSON payload.

You can test the endpoint using the following curl command

curl --location --request POST 'localhost:5000/v1/predict-price' \
--header 'Content-Type: application/json' \
--data-raw '{
    "latitude": 38.0100,
    "longitude": -122.64,
    "housing_median_age": 36.0,
    "total_rooms": 1336,
    "total_bedrooms": 258,
    "population": 678,
    "households": 249,
    "median_income": 5.5789,
    "ocean_proximity": "NEAR OCEAN"
}'

You should get the following response

{"predicted_house_price": 312116.03}

Production

You are encouraged to use Docker for building and running the application in a container in the production environment. To run the app, use docker-compose

docker-compose up

Application should now run on http://localhost:5000/.

To change the port (default is 5000), edit the docker-compose.yml file.

# ...
    ports:
      - 5000:80     # replace 5000 with your desired port number
# ...

The application is deployment ready. It is served using uWSGI and nginx for best production performance. If you need to tweak the production parameters like number of processes (4 by default) edit the .docker/uwsgi.ini file (at your own risk).

Development

Setup

Preferably create the Python evironment first. You need to do this initial step only once.

python3 -m venv venv

Then activate the environment

. venv/bin/activate

Install dependencies

pip install -r requirements.txt

Install the application package

pip install -e .

Unzip the pickled model file

tar -zxvf data/model.tgz

Usage

Set up the env variables

export FLASK_ENV=development
export FLASK_APP=housing_prices

Run the application

flask run

Application should now run on http://localhost:5000/.

API Documentation

See the Swagger documentation with executable examples for the whole API on http://localhost:5000/v1/apidocs/.

Testing

There is a suite of tests stored as a Postman Collection in the tests/postman_collection.json file. Import it into Postman and hit Run collection to see the results.