Using D3 and the IPython Notebook

Monte Lunacek




slides:http://bit.ly/1hUqI4W
code: http://bit.ly/1bbqyVp

Outline

• Motivation
• Ways to use D3 in the Notebook.
• Examples
• Conclude
• Download and try it!

Creating a Visualization

Process

• Raw data
• Processing
• Exporatory Data Analysis
• Analysis
• Visualization

Methods: R, Python, Matlab, IDL, Excel,...

Python: notebook, pandas, scikit-learn

What is the IPython Notebook?

• Web-based, interactive Python environment.
• Text, Latex, Markdown
• Embeded graphics
• Remote computing
• Process documention
• Lots of ways to share:
  • nbviewer.ipython.org/
  • HTML, Latex, PDF
  • Slides

Why D3 in a Notebook?

• Interactive, web-based, beautiful
• Kernel: do almost anything in 2D.
• Visualization not available, not interactive
• Parallel Coordinates
• Crossfilter
• Cubism
• A single environment
  • Leverage modern web technology and
  • computation.

2012 PyCon video
http://www.youtube.com/watch?v=F4rFuIb1Ie4
D3 at ~40:00 minutes

Landscape of D3 and IPython Notebook

And many more...

Custom D3 visualization in a Notebook

• Start with a stand-alone HTML5 visualization
• Create a template with Jinja2.
  • Control variables (e.g svg height and width)
  • Control data with numpy, pandas
• Load d3.js
• Render the template in the Notebook.

Conclusions

• IPython Notebook is a satisfying way to work with data.
• D3 is a flexible, interactive, and web-based.
• Combining them is straight-forward:
  • Jinja2 template
  • Data with numpy and pandas
• The Notebook can capture your process.
• Easy to share.