econgineer
econgineer

Reputation: 1137

Sharing Interactive IPython/Jupyter Notebooks for Non-Technical Users

Is there an optimal way to share/serve a readonly-ish Jupyter notebooks with the input code cells hidden, but connected to a live kernel so users can interact with widgets?

I understand there are ways to embed data for widgets, but often the widgets require rerunning a query against a larger data sets.

I've recently started learning to use widgets and they'd be really useful to share interactive data with our non-technical staff, but I haven't figured out a good way to do so yet.

Having this capability would also reduce the desire for expensive tools like Tableau.

Upvotes: 5

Views: 2475

Answers (2)

moon
moon

Reputation: 166

Take a look Apache Zeppelin https://zeppelin.apache.org. Which is closest opensource software that you can share a notebook with code hidden and let other users run it. Apache Zeppelin has Dynamic form, which is not exactly widgets but used for the same purpose. Zeppelin supports notebook level access control so you can achieve what you want.(disclaimer: I'm creator of Apache Zeppelin)

You can also take a look notebook sharing service https://www.zepl.com. Which enables

  • Access control on the notebook
  • Notebook updates auto synced from Github / S3
  • Report looknfeel (Hide code for non-technical people)
  • Comments on the cells
  • Team and Workspace to organize sharing
  • Both Jupyter notebook and Zeppelin notebook format support

This can be an possible alternative on sharing your notebook with non-technical users. (disclaimer: I'm co-founder of ZEPL)

Upvotes: 0

Eoin Murray
Eoin Murray

Reputation: 1955

If you use kyso.io (disclaimer: I founded kyso) you can share notebooks online easily:

  • It will you can show/hide code so non-technical people can read it
  • Interactive visualisations work perfectly, like bokeh and plotly.

Heres some examples:

you can share a notebook using these commands

kyso create "some study title"
kyso push "a version message"

Currently the notebooks are static renderings, but soon we are going to host runnable Jupyter dashboards.

Upvotes: 4

Related Questions