house-price-predictor-service

Python service that is able to fit a predictor model and deliver predictions and recommendations about the prices of houses in the US.

To consider:

• the model used for prediction was trained using the notebook data_exploration.ipynb
• the model is saved in the folder data as pickle file: house_price_model.pickle
• the application service was implemented using FastAPI and server with uvicorn.
• the service can be configured using the config.py file. Here you can edit the host and port used to run the app.
• the endpoints in the app can be tested using the notebook test_service.ipynb
• automatically FastAPI enable the app documentation in the url [host]:[port]/docs
• some unit testing functions are in the file test/test_service.py. Run it using python -m pytest

Steps to start the service:

• create a virtual environment
• enable the previous virtual environment created
• install the requirements using the requirements.txt file: pip install -r requirements.txt
• run the app: python main.py