/Kaggle-Titanic

Intro into doing a kaggle competition with all the necessary data science tools and techniques.

Primary LanguageJupyter Notebook

Kaggle Titanic Competition

This is the dev branch of the repo aimed at modelling and making submissions for the Kaggle Titanic Competition with the code from Juputer Notebooks. This repo will be updated as I try better models to improve and get a higher score.

To get started:

  1. Clone this repo on your local machine
  2. At the terminal, run:
pipenv shell
  1. Then install the requirements:
$ pipenv install -r requirements.txt
  1. Finally, open the jupyter lab environment:
pipenv run jupyter lab

The notebooks

  1. First submission- titanic-pred-raw.ipynb, uses linear SVM model to predict survivors aboard the Titanic. Accuracies gotten:
  • validation: 83.8%
  • test(submitted): 77.2%