This repo is a proposed solution to the "Data Engineering Challenge" first step on the Titanic kaggle competition.

Setup

First copy the .env.sample file to .env and fill in the values.

cp .env.sample .env

Then download your service account key from Google Cloud Platform and save it as credentials/service-account.json.

Once this is done and if you are using pyenv-virtualenv you can run the following command to setup the project.

make init_env

Running the pipeline

To run the pipeline you can use the following command:

make train

Web application and API

Launch the api in local with :

make run_api

You should test it by following the link that will be displayed in the terminal. Don't hesitate to go the /docs endpoint to see the documentation of the API. You can also test the API with the make test_api rule in another terminal.

Finally launch the streamlit app with :

make streamlit