Project Summary

• This project is to build an API application that can be used to infer whether a person makes over 50K a year based on census data. The API is built using FastAPI and deployed on Heroku. The model is trained using data from the census income dataset.

Environment Setup

• Create a virtual environment CONDA CREATE -n <env_name> python=3.8
• Activate the virtual environment CONDA ACTIVATE <env_name>
• Install the required packages PIP INSTALL -r requirements.txt
• Run the application locally with localhost and port as 8000 uvicorn main:app --reload
• Run the tests pytest -c test/pytest.ini

Model Information

• The model card is in the model_card.md file.

Project Structure

├── data
│   ├── census_income.csv
├── model
│   ├── lb.joblib
│   ├── encoder.joblib
├── src
│   ├── ml
│   │   ├── model.py
│   │   ├── data.py
│   ├── train_model.py
│   ├── get_inference_fromAPI.py
├── test
│   ├── pytest.ini
│   ├── test_main.py
│   ├── test_ml.py
├── Procfile
├── Aptfile
├── main.py
├── requirements.txt
├── model_card.md
├── README.md
├── LICENSE.txt
├── outputs
│   ├── example.png
│   ├── live_post.png
│   ├── continuous_deployment.png
│   ├── live_get.png
│   ├── slice_output.txt