Over fifteen years ago the Agile Manifesto was created for the express purpose of developing better software. Yet better software is not the underlying *reason* that organizations hired all those agile teams. The software was the “means” and greater organizational value was the “ends.” Now many of these same organizations are looking to capitalize on a new resource. They’re collecting petabytes of structured, semi-structured and unstructured data. Exploiting this digital raw material has many of the same challenges as software development. That’s why many well-formed **data science teams** struggle with some of these same questions:
* Are we creating something valuable?
* Can we closely coordinate with our customer?
* How can we quickly pivot to take advantage of unexpected outcomes?
Many long established agile team practices could also apply to newer data science teams. These teams require a lightweight empirical framework to help deliver products of pure discovery. The core difference is the iterative product. Instead of minimum viable software, these teams will deliver frequent valuable insights.
This talk will show how to apply a lightweight agile framework to data science teams. These teams can use modified version of common agile practices such as user stories, cross-functional teams and frequent iterative delivery.