The Data Analysis Tools Series at the Berkeley Institute for Data Science

IPython - Omoju Miller

February 25, 2015 at 5-6:30pm in BIDS, 190 Doe Library

Attending

I counted 35 people. These included, at least:

Omoju Miller

Omoju Miller is a PhD candidate at the University of California at Berkeley researching artificial intelligence. She is also a software technologist, start-up advisor, and educator.

IPython

IPython is an interactive interpreter for programming with Python (and now many other languages).

Easy, Peasy, Lemon Squeezy

Omoju suggests that, to work, teach, or collaborate, development tools need to be as easy as possible to install and use.

Things that she mentioned in this regard:

The IPython Notebook

To start up the ipython notebook, crack open your terminal and type:

ipython notebook

That starts up a server which serves ipython notebookes (usually to localhost:8888 or similar). This command, therefore, will automatically open a browser instance with a view of your directory. This will allow you to open up any ipython notebooks in that directory. It also allows you (with a button) to create a new notebook in that directory.

The Oscars

Omoju showed an example from Mining the Social Web about using the twitter api. She demonstrated how she was able to use the ipython notebook to access the twitter firehose and filter out tweets concerning the Academy awards.

LaTeX in Markdown cells

She demonstrated also how to include LaTeX in a markdown cell. First, create a markdown cell, then include math:

Courtesy of MathJax, you can include mathematical expressions both inline:
$e^{i\pi} + 1 = 0$  and displayed:

$$e^x=\sum_{i=0}^\infty \frac{1}{i!}x^i$$

Using IPython Notebooks with GitHub

A troublesome issue with IPython notebooks is the extra information that is held in the json. That doesn’t version control as beautifully as plain text.

To avoid extra headaches, clear all output cells (Cell - > All Output - > Clear) before you commit your ipython notebooks.

Magics in the Notebooks

Omoju describes these as Development Powertools. Magics are special tools and functions. They are often preceeded by one or more percent signs (%). Some examples:

Plotly Notebook Examples

Plotly is a plotting tool. On their websites, there are some fun examples in the gallery of IPython notebooks .

Use Cases

Omoju suggests using IPython notebooks for lots of stuff, including:

Learning more with Books

Omoju recommends The Ipython Cookbook.

Lightning Talks

Martin Magdinier : OpenRefine

OpenRefine is a tool for helping to clean and process data. You can do this with very limited data processing skills, but it is also useful for more skilled analysts.

Refine runs a local java-based server on your local machine and it opens up a browser instance to provide an interface for loading data, parsing it, identifying close duplicate categories, cleaning it up, and exploring it somewhat with filters and views.

Brian Hamlin : Geospatial stuff!

Geospatial data is everywhere. An example is found at www.pism-docs.org

OSGeo-Live

Daniel Wooten : Prompt Magic

Dan Wooten uses SSH to get to computers all over the place. He likes to be able to tell what computer he is on simply by the color of the prompt in his terminal.

In the .bashrc file, you can export a modified PS1 variable to change the content and color of your prompt.

BUT, the code necesary for specifying the right thing is pretty hideous. Let someone else figure out the syntax for you with : PROMPT MAGIC!!!

Share