Deploy-Loan-Approval-Classifier-on-Azure-cloud

In this project, I'll walk you through deploying any Machine Learning project, Loan Approval Classifier model in our case, on Azure Cloud.

Directory Structure

• Data: Data used for this project. However Train/DataPrep.py will be used to import real time data from Database for our this project.

• configs: It contains the configuration files that will be utilized to store different credential data such DB related credentials, etc.

  • configs/config.yaml: Stores DB server's credentials.
  • configs/env.yml : Dependencies that need to be installed for this project.

• Train: It contains python scripts that will be utilized for Model training.

  • Train/DataPrep.py : Import Data from DB and preprocess it.
  • Train/model.py : Python script for model training and export pickle file to outputs directory.

• outputs : Stores various output and pickle files such as Encode,pre-processor used in training step, Trained Model,etc.

• Prediction: It contains python scripts that will be utilized to predit test, or actual data.

  • prepare.py: Preprocess and predict response variable for test data
  • score.py: Azure script file.

• Performance: It contains python scripts that will be utilized to assess model performance on actual data and performance across different run_Ids over the time.

  • FindActuals.py: Calculate model performance and store logs in Database.
  • Performance/modelPerformance.py: Calculate model performance over the time across different run_Ids to check model or data drift.

• utils.py: It contains different utility functions.

• main.py: Python script to define order to follow to execute different relavant Azure experiment scripts.

• requirements.txt: Python dependencies that will be required to be installed to run this project.

Instructions for Installation:

Dependencies:

• joblib>= 0.16
• scikit-learn>=0.23
• pandas>= 1.0
• numpy>=1.1
• pyodbc==4.0.30
• urllib3==1.25.9
• sqlalchemy==1.3.18
• pyyaml>=5.3
• azureml-sdk>=1.18.0

License:

This project is open-source and distributed under the MIT License. Feel free to use and modify the code as needed.

Issues:

If you encounter any issues or have suggestions for improvement, please open an issue in the Issues section of this repository.

Contact:

The code has been tested on Windows system. It should work well on other distributions but has not yet been tested. In case of any issue with installation or otherwise, please contact me on Linkedin

About Me:

I’m a seasoned Data Scientist and founder of TowardsMachineLearning.Org. I've worked on various Machine Learning, NLP, and cutting-edge deep learning frameworks to solve numerous business problems.