Wakari Sharing

 

 

You can explore the plots (panning and selecting) above, and I’ve shared the code which generated them here. In addition to viewing the notebook, you can use the code in Wakari by clicking on Add this notebook to my Wakari Account. I can continue building out my analysis of NOAA weather data and everyone can both view and run my code — verifying my methodology and reproducing my results.

Use Cases

This feature will obviously be useful for education and Academia. Wakari provides a consistent back-end, so code which runs in one user space will run in all user spaces — producing the same results everytime. No more OS and library problems.

Because Wakari is a generalized data analysis environment, the sharing and reproducibility problem is solved for many fields other than Academia like Finance. Quants and Business Intelligence engineers can develop strategies and analyses and easily collaborate with one another without passing excel files around the office and without the headache of managing environments.

Creation and Re-Creation

Humans are driven by newness at a fundamental level. But newness on its own doesn’t get us very far, especially in the sciences; instead we need also to be able to reproduce and verify exciting new information. The desire to re-create — a play, a house, a schematic, a ‘hello world’ code snippet — is perhaps as innate as the original desire to create.

The sharing of information may enhance one’s ability to re-create/reproduce an experiment or analysis but it doesn’t guarantee it. In academic writing, authors are usually expected to provide their methodology along with information about how the reader can reproduce their work. But all too often, the reproducibility of academic texts falls short. Recently, members within a variety of academic fields have called for new standards in reproducible research. One particularly important aspect of this reproducibility involves the sharing of source codes. In one Nature article the authors write:

We argue that, with some exceptions, anything less than the release of source programs is intolerable for results that depend on computation.

Future work

Currently, shareable notebooks are limited to only public viewing. We are working hard to implement group permission for notebooks, folders, etc. as well as premium features to better support different levels of access to code, data, plots, and environments. For more information please go to www.wakari.io and signup for a free beta Accountnt now.


About the Author

Ben Zaitlen

Data Scientist

Ben Zaitlen has been with the Anaconda Global Inc. team for over 5 years.

Read more

Join the Disucssion