05satyam/risk_prediction_loan_application

Risk Prediction for customer loan application

Project Structure

project_root/
│
├── src/
│   ├── pickel
│   │       └── model.pkl
│   ├── eda_notebooks
│   │         └── exploratory_data_analysis_initial_phase.ipynb
│   ├── base_model.py
│   ├── data_ingestion.py
│   ├── data_preprocessing.py
│   ├── model_deployment.py
│   ├── model_evaluation.py
│   ├── model_training.py
│   ├── pipeline.py
│   ├── utils.py
│   ├── pipeline.ipynb
├── .env
├── .gitignore
├── config.json
├── app.py
├── requirements.txt
├── index.html
├── .streamlit
│       └── config.toml
├── pipeline_stages.png
├── data/
│     ├── loan.csv
│     ├── plots
│           ├── corelation_btwn_principal_apr_repayment_amt.png
│           ├── correlation_heatmap.png
│           ├── loan_application_status_with_repay_trends_bargraph.png
│           ├── relation_between_cleint_paying_freq_and_fraud_score.png
│           ├── turnover_clients.png
│       
└── project_output_images/
            ├── 1.png, 2.png, 3.png, 4.png

Steps to run the project

• brew install libomp {install as per your operating system- i was using macos}

• To run the application

  • Go to root folder {project_root}
  • run: pip install -r requirements.txt
  • run: streamlit run app.py

• Now If you want to run the piepline script, Please open pipeline.py and uncomment the main and run the script

Deatuils on Project structure:

• src/ : Contains all the Python scripts for data ingestion, preprocessing, model training, evaluation, deployment, pipeline(main class) utils, pickel file for model, EDA notebook.

  • pickel/: Directory to contains the saved (trained model)
  • notebooks/ : Directory to contains the Exploratory Data Analysis(EDA) notebook to evaluate and understand data
  • app.py - The Streamlit application script for the web interface.
  • utils.py - Helper script to load env variables and get file paths for csv and model path
  • pipeline,ipynb - a notebook to run the and test pipeline block by block

• .env : Set the environemnt variables for different paths

• config.py : Helper script to store

• Which columns to remove during training {columns which are not at all needed}

• Store the target column name

• requirements.txt - Lists all the Python packages that need to be installed.

• data/ - Directory for storing input data and any processed datasets and plots generated during .ipynb notebook evaluations.

• project_output_images/ - This folder contains my output images of the web-ui and display outputs of running model following pipeline stages

• index.html : documentation html for intermal working of the pipeline

Details on Pipeline:

• Read index.html

Possible error while setting up

• Streamlit axios error
  • Resolve:
    • Link: https://discuss.streamlit.io/t/axioserror-request-failed-with-status-code-403/38112
    • Created a new folder called .streamlit in root path and create a “config.toml” file inside it.
    • Inside config.toml

      [server]
       enableXsrfProtection = false
       enableCORS = false