Login Sign up

Monday Oct. 12, 2020, 4 p.m.–Oct. 12, 2020, 4:30 p.m. in Jupyter Community: Practices

Jupyter as an Enterprise “Do It Yourself” (DIY) Analytic Platform

Dave Stuart

Audience level:

Brief Summary

Jupyter’s use as an Enterprise “Do It Yourself” Platform puts the power of analytic development and data science capability directly in the hands of the business analysts closest to the analytic challenges. This real world use-case describes the success as well as the cultural and technical challenges from growing a community of more than 12,000 Jupyter users within a single enterprise setting.


In this case study from inside the US Intelligence Community (IC), we details how Jupyter has empowered thousands of business analysts to create their own Do It Yourself (DIY) analytic solutions. Five years of concerted effort to evangelize Python and Jupyter within this large enterprise setting have netted tremendous gains. Through the right combination of outreach and training, alongside platform enhancements, business analysts finally find themselves on the same side of the wall as solutions development. Jupyter has empowered this community of analysts not traditionally steeped in technical disciplines like software engineering - to translate their tradecraft into code, making that tradecraft more reproducible and more efficient.

But the story doesn't end there – a DIY analytics movement introduces new challenges, including an abundance and redundancy of solutions. With two thousand Python authors, and twelve thousand Jupyter users, this movement would fail under its own weight without significant efforts to manage, curate, sustain, and provide a corporate “path to production” for the hardest-hitting new capabilities.

Our talk will describe this path to Jupyter adoption from the vantage point of the enabling team, what challenges we faced (anticipated and unanticipated), and how we overcame them to transform business analysis in the IC. We will detail the tools and approaches we have developed to manage, curate, and sustain crowd-sourced development of Jupyter notebook based analytics. We’ll look at the training paths used to introduce Python and Jupyter into communities that most often lack prior backgrounds in coding. And we will outline the work we have done to lower the barrier to entry through the adoption of solutions like Voila Dashboards that help make this platform as approachable as possible.