Airbnb open sources data-science-sharing platform

0 Posted by - 3rd November 2016 - Technology

Most organizations have well established procedures for vetting and sharing computer code. But what about data analysis?

Important findings are often held in “a mixed bag of presentations, emails, and Google Docs,” two members of Airbnb’s engineering and data science team blogged at Medium in February. When someone in the organization wants to locate and use that existing work, they often have to track down updated code and waste time checking and reproducing earlier results. And then they’ll typically distribute their own findings “through a presentation, email, or Google Doc, perpetuating the cycle.”

After considering various ideas on how to solve this problem, Airbnb created an internal Knowledge Repo, combining git version control and Markdown templates for reporting results. Airbnb recently open-sourced its Knowledge Repository Beta, seeking contributors to help move the project forward.

Git allows the same sort of peer review and version control that developers typically use to collaborate on code, while Markdown offers a mixture of text and code in a single, easily reproducible file. You can see RStudio’s tutorial on R Markdown for more info of what Markdown in general can do. Markdown is available for other languages such as Python as well.

The Airbnb framework setup requires Python and supports “knowledge posts” in several formats. via #CIO, #Technology