Science Docker

This is a Docker Compose configuration built to do some data science and machine learning research. It features the most popular Python libraries.

The image is based on Debian Jessie distro.

Requirements

• Docker
• Docker Compose

Instructions

• Clone the repository
• Build the container
• Launch the container

Features

Features

Math, data manipulation

• Numpy
• Pandas

Web scraping

• Beautiful Soup
• Request

NLP (Natural Language Processing)

• Gensim
• NLTK
• Pyenchant
• Compound word splitter
• Spacy

Visualization

• Matplotlib
• Seaborn

Deep Learning

• Tensorflow
• Keras

Jupyter Notebook

• Make sure port 8888 is available on your host machine
• Upon launching the container, visit http://localhost:8888
• Use science as password
• If you want to change password or do other customizations, you can find a jupyter_notebook_config.py configuration under the docker/ directory

Work dir

• A volume under /work is mounted as shared work folder between host and container

Todo's

• Refactor container to use a lighter image (Alpine?)
• Try to implement and improve configurability

Contribution

Please feel free to PR this repository and contribute. This repository was born tailor made on my research and learning needs but it would be nice to make it a tool for everyone, maybe with many customization options.