Breaking Data Science Open

Chapter 9. Data Science Collaboration

We have previously discussed the data scientist unicorn—that mythical creature that has deep computer science, advanced analytics, and business or domain expertise. If this person exists at all, they are not found in the typical enterprise setting. More commonly, the various perspectives and skills required to deploy high-value data science applications are supplied through the aggregated skill and knowledge of many people in an organization. In the past, collaboration meant that teams operated as though they were in a relay race, handing off the baton once their lap of the track was complete. The adoption of Agile methodologies into the workforce means collaboration looks more like a soccer team, working together to get the ball down the field and score the goal. While soccer players have the luxury of seeing, hearing, and interacting with each other on the field while trying to score the goal, however, most data science teams span departments and locations, making it much more difficult to develop an analysis workflow, perform an ad hoc analysis, or create a data science application. Data science teams need to share expertise and data across these organizational and geographic boundaries to build, test, and deploy data science models that drive value for their organization. When a data science team is empowered to work together as a tight ensemble, helping each other and boosting each other’s efforts, they can deploy better-performing data ...