Kaggle Titanic Challenge

The Goal of this project is to predict the survival of the person given the information.

Pre-requests

• Python 2.7
• Pandas: data analysis
• Seaborn: data visualizatoin
• scikit-learn: random forest
• Keras: neural networks
• Theano: tenser operations, used as Keras's backend

Quick start

• See visualize/Data_Explore_Jia.ipynb for explorations over data and features.
• Add <repo path> to your PYTHONPATH if you want to import this as package.

What's new

• Restructured as package format. Run

cd src/test
python testdata.py

to see if data submodule works.

Structure

File structure is like this:

./
|- titanic/ # all reuseable sources as a package
  |- __init__.py
  |- data/ # modules about data parsing and processing
    |- __init__.py
    |- dataset.py 
    ...
  |- nn/ # modules about NN
    |- __init__.py
    |- model.py
    ...
|- data/ # dataset and processed data
  |- original/
    |- xxx.csv
    ...
|- visualize
  |- xxx.ipynb
|- README.md
...