Doing Data Science the Agile Way

Working as a Data Science consultant over the years, I had the privilege of collaborating with numerous and diverse software engineering teams, each following its own software development practices.

Every time I joined a new team I had to adjust to those practices; this meant I had to do Data Science in a way that would adhere to general software development principles while aligning with team values and planning conventions. It wasn’t always easy to do so.

In this talk, I attempt to provide a guide on how to practice Data Science as part of an Agile team:

• What is a Data Scientist’s role in an Agile team?;
• the importance of setting clear expectations to stakeholders;
• how experiments (not tasks) can be done in a way that always adds to the team’s understanding;
• how to extract value from intermediate results and incomplete solutions//models;
• how should work be split?
• knowing when your work is done.

You can view Christos’ slides below:

You can view Christos’ presentation below: