Skip to content

An End to End Azure ML pipeline, predicting Adult income category from census data.

Notifications You must be signed in to change notification settings

rajtulluri/End-to-End-Azure-ML-pipeline

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

End-to-End-Azure-ML-pipeline

A sample End to End Azure ML pipeline deployed on heroku.

ML studio experiment and Web service

The Machine Learning model built on Azure ML studio, predicts the Adult income category (>50k or <=50k) using a tuned and validated Two-class boosted Decision tree model. The trained model is converted to an Azure web service.

Two images below are the ML studio pipelines

The training experiment

Training experiment

The predictive experiment (web service)

Web service

The Web service offered by Azure is a paid service, hence the API no longer exists. Please create your own Azure Web service based on the image and save the URL and API key in the respective files under resources folder.

Dash app (python)

A Dash app is coded to post requests to the Azure web service created. The dash app is deployed on Heroku. The file structure of the app is,

__ app.py
__ Layouts
  |_ layout.py
  |_ tab1_layout.py
  |_ tab2_layout.py
__ requirements.txt
__ Procfile

The app file contains the main server code along with callback functions for dynamic API calls. The Layouts folder contains various layout files for the app to render.

Heroku deployment

Install the following dependencies in the env using pip

dash
pandas
numpy
gunicorn
flask
urllib

Use the following command

pip install _package_

Create a file named Procfile and fill its contents with,

web: gunicorn app:server

The Procfile has the instance of the server which the heroku platform will run Create a new file requirements.txt with a list of all dependencies,

pip freeze > requirements.txt

Run the following commands in succession to deploy the app on heroku

heroku create my_app
git add .
git commit -m "Initial push"
git push heroku master
heroku ps:scale web=1

The heroku server will provide the deployed application's link. The app above can be viewed at - https://azure-ml-pipeline.herokuapp.com/

If using a conda environment, then before the above steps, execute the following Create a conda env

conda create -n env_name python=3.6

Activate the env using,

conda activate env_name

Continue with the steps mentioned above.

NOTE:- Build a personal API using the Azure ML studio. Since it is a paid service, the API invovled in the app is no longer active.

Releases

No releases published

Packages

No packages published

Languages