A simple Flask application demonstrating how to deploy and access a trained Machine Learning model.
See Machine Learning Handbook for details about the training process.
The /predict
endpoint expects a POST request whose body must contain the input data for a housing district (population
, median_income
...) as JSON. See an example of valid JSON payload.
It returns the model's prediction (median price for the housing district) as a JSON object with a median_housing_price
property.
Launch the following command to run the server locally.
> env FLASK_APP=main.py FLASK_ENV=development flask run
You must also update the JavaScript prediction_url
variable in templates.index.html
to point to the local web server instead of the Heroku app.
// Switch between local and remote (Heroku) deployment
// const prediction_url = "http://127.0.0.1:5000/predict";
const prediction_url = "https://housing-prices-api.herokuapp.com/predict";