Usage of the Jupyter Notebook in Publishing



In [1]:

    
%load_ext load_style
%load_style images.css
from IPython.display import display, Image, HTML

The Notebook is being used to publish content that contains code+data to a wide range of audiences.

LIGO's gravitational wave discovery

In early 2016, the Laser Interferometer Gravitational-Wave Observatory (LIGO) announced the observation of gravitational waves.



In [2]:

    
Image('images/ligo-press.png', width='80%')









    Out[2]:

Along with this announcement, they release all of their data along with a full analysis as Jupyter Notebooks.

One of our developers (@minrk) created a binder to allow anyone to reproduce their analysis (https://github.com/minrk/ligo-binder).

O'Reilly books

O'Reilly media has been publishing books that are accompanied by Jupyter notebooks. Here are a few examples:

http://shop.oreilly.com/product/0636920034919.do?sortby=publicationDate

https://github.com/jakevdp/PythonDataScienceHandbook



In [10]:

    
Image("images/python-data-science-handbook.png")









    Out[10]:

http://shop.oreilly.com/product/0636920030515.do

https://github.com/amueller/introduction_to_ml_with_python



In [11]:

    
Image("images/intro-ml-python.png")









    Out[11]:

http://shop.oreilly.com/product/0636920030195.do

https://github.com/ptwobrussell/Mining-the-Social-Web-2nd-Edition



In [3]:

    
Image("images/oreilly-socialweb.png", width="80%")









    Out[3]:

Python for Signal Processing

Jose Unpingco has written a series of blog posts on Signal Processing Jupyter+IPython Notebooks. These blog posts were the basis of a full length book Python for Signal Processing published by Springer in 2013.



In [4]:

    
Image("images/p4sp-clipped.png", width="60%")









    Out[4]:

O'Reilly Media's website and project Thebe

O'Reilly Media is embracing the Jupyter architecture and notebook at a deep level. The "learning" part of their website not includes content with embedded code cells that allow readers to run code and experiment.

https://www.oreilly.com/learning

This is done using Thebe, an open-source project from O'Reilly that enables the embedding a live code cells on any web page:

https://github.com/oreillymedia/thebe



In [6]:

    
Image("images/oreilly-beta.png", width="80%")









    Out[6]:

More recently, O'Reilly has launched Oriole Online Tutorials that combine live code with follow-along video:

http://www.oreilly.com/oriole/index.html



In [7]:

    
Image("images/oreilly-oriole.png", width="80%")









    Out[7]:

Reproducible academic publications

Publications in traditional academic journals are being accompanied by notebooks that reproduce and expand upon the main computational results of the work. See this section of the Notebook Gallery for a full list.

Here is one notable example from Nature Genetics by Gross et al., titled Multi-tiered genomic analysis of head and neck cancer ties TP53 mutation to 3p loss.



In [8]:

    
Image("images/nature-genetics.png", width="80%")









    Out[8]:

The authors published a GitHub repository with notebooks that replicate the results:

https://nbviewer.jupyter.org/github/theandygross/TCGA/tree/master/Analysis_Notebooks

What if?

In 2010 Reinhart and Rogoff published a paper titled "Growth in a time of debt" in the American Economic Review. The work was highly regarded and used as justification for austerity mearsures in the years following the 2007-2008 economic recession.

In 2013, Herndon, Ash and Pollin published a critique showing that the computational work in the original article had errors and that the results were not reproducible.

Vincent Arel-Bundock has replicated the analysis of Herndon, Ash and Pollin as a Jupyter notebook.

http://nbviewer.ipython.org/github/vincentarelbundock/Reinhart-Rogoff/blob/master/reinhart-rogoff.ipynb

What if Reinhart and Rogoff had distributed their original analysis as a notebook?



In [9]:

    
Image("images/reinhart-rogoff-replication.png", width="80%")









    Out[9]: