Machine Learning Web App Using Flask

This web app provides a simple summary of using a variety of machine learning models (Sklearn, Tensorflow, and Azure Machine Learning) deployed to a python web app (Flask)

• Web App Link: https://python-ml-web-app-orthopedics.herokuapp.com/

How to Use

• Clone this repository with: git clone https://github.com/van-william/flask_app_ml.git
• Reference the sample notebooks in the notebooks directory
• Leverage the simple web framework (html + css) for your own web app projects

Web App Hosting

• This project is hosted on Heroku: https://python-ml-web-app-orthopedics.herokuapp.com/
• The repository contains context on the necessary Profile, runtime.txt, and other required files for Heroku hosting

Model Deployment

• Originally, Microsoft Azure Machine Learning was used for model deployment (Jupyter Notebook and Azure Machine Learning API code are still in the repository), but the primary predictor was simplified to be a localized .pkl file from scitkit-learn
• There is a notebook for Deep Learning (Tensorflow), but the library was too big for the free version of Heroku. However, the machine learning code is still referenced

Architecture Overview

References

• I leveraged this detailed Flask tutorial for help in building the web page
• There are several examples for deployed machine learning models
  • https://www.geeksforgeeks.org/deploy-machine-learning-model-using-flask/
  • https://towardsdatascience.com/deploy-a-machine-learning-model-using-flask-da580f84e60c
• Francois Chollet has a good repository of deep learning examples for reference as well
• Data Source Citation: Dua, D. and Graff, C. (2019). UCI Machine Learning Repository [http://archive.ics.uci.edu/ml]. Irvine, CA: University of California, School of Information and Computer Science.