Data Product Development and Deployment with Streamlit

Requirements

1. Setup Github account
2. Create virtual environment

python -m venv venv

3. Activate virtual environment

• Windows

venv\Scripts\activate

• Linux/MacOS

source venv/bin/activate

4. Install required packages:

pip install -r requirements.txt

Prelude: Try Streamlit

1. Create toy application with Streamlit.
2. Push repository to GitHub.
3. Deploy on Streamlit community cloud.

Sample application code: toy-app.py

Step 1: Train and Save Model

1. Perform EDA and model development on Jupyter notebook.
2. Develop training and model registry scripts to automate model training and persistance respectively.
3. Run the training script to train a loan approval model:

python src/training.py --data_path data/loan_dataset.csv --model_path models/ --f1_criteria 0.6

Sample model training notebook: DSSI_LoanModel.ipynb
Sample training script: training.py
Sample model registry script: model_registry.py

Step 2: Create App and Load Model

1. Develop an inference script to serve predictions.
2. Create a loan approval application with Streamlit that automates decisions with user inputs and trained model.

Sample inference script: inference.py
Sample application code: app.py

Step 3: Test App Locally

Run and test the application locally:

streamlit run app.py

Step 4: Deploy App Online

1. Commit repository to GitHub.
2. Deploy on Streamlit community cloud.