sales_prediction_api's Introduction

Deploy a ML model using Flask

Based on: https://github.com/abhinavsagar/machine-learning-deployment

The app deploys a simple machine learning model to predict sales in the third month based on the sales in the first two months.

Run the app and make a query

With Docker

Run

$ docker run -dp 8080:8080 staeff/salespredictionapi

open http://localhost:8080/ in a browser or make a curl request:

$ curl -X POST -H "Content-Type: application/json" -d '{"rate":5, "sales_in_first_month":200,"sales_in_second_month":400}' http://localhost:8080/results

Without Docker

$ python3 -m venv .venv; source ./.venv/bin/activate
(.venv) $ pip install -r requirements.txt
(.venv) $ python app.py
# in another terminal window
(.venv) $ python request.py
523.8530977224776

When the flask process is running you can also open http://127.0.0.1:5000/ and enter data into the website.

or use curl:

$ curl -X POST -H "Content-Type: application/json" -d '{"rate":5, "sales_in_first_month":200,"sales_in_second_month":400}' http://localhost:5000/results

Repo contents

app.py — a Flask app that receives inputs through website or API call, computes the predicted value based on our model and returns it.
HTML/CSS — frontend code for the flask app
request.py — makes an example API call
model.py — computes a linear regression model to predict sales in the third month based on the sales in the first two months. No need to run this. The model is already stored in the repo.

Recommend Projects