Abstract:
More and more often, data science capabilities need to be developed and integrated with software systems to deliver the intended value to the organization. While most data science practices come from scientific and/or academic domains, most software engineering projects are executed using methods and practices developed in industry. Executing data science development along-side, and complimentary to software engineering has proven a challenge in many organizations. In this short tutorial, we will cover how the basic components of agile software development can be adapted to data science efforts. We will discuss how to define “stories” and “epics” for data science, how to manage and prioritize backlogs, stand-ups, sprint reviews, and how to communicate with stakeholders. The methods covered have, in our application of them, resulted in better-managed stakeholder expectations, kept teams focussed on delivering a business capability, and helped to ensure that we deliver the intended impact to the business.
Presenter:
John Akred, Silicon Valley Data Science
