Notebook Notebook:

• What we just saw is awesome as an interactive shell, but not the best for:

  1. Saving/Reusing Work:

     • Don't want to have to keep a session open forever.
     • No good way to share a session.
     • %edit helps for writing larger chunks of code.

  2. Minimal support for rich display.

  3. Sometimes you want to include other content besides Python code (e.g. instructions, annotations, diagrams...).

• Quit IPython, run ipython notebook.

  • This runs a notebook server, which I can connect to by pointing my browser (or another application that knows how to read the format) to localhost on the appropriate port (default is 8888).

Pandas Notebook

Zipline Notebook

Research Env Sneak Peek:

• We've seen that IPython Notebook acts as a nice middle ground between the static nature of a script/library and the dynamic (but impermanent) nature of a shell session.

• We've also seen that IPython Notebook enables sharing computation, which includes not just Python code but also data visualizations, documentation, and other content.

• And we've seen that we can use IPython Notebook with the state of the art tools (like pandas and zipline) to get even more out of these tools.

• Said at the start that we're thinking now about how to make the research process on Quantopian better...

  • One of the most powerful features of notebook is the fact that it you don't actually need to be running the notebook server to connect to and edit a notebook.

  • We want to make it possible for people to run notebooks on our computing environment in a way that can still take advantage of all awesome data sources and APIs that are currently only available on Quantopian.

• get_pricing notebook.

• get_backtests notebook.

• local_csv notebook.

• This project is in a very early alpha, but we're very excited about the possibilities this opens up. We're already starting to use the idea of remote notebooks in our internal research and operations efforts.