35  Session 1: Design Thinking in Data Science

This week explores the question “Can we approach data science as a design problem?” and discusses how one can embrace a user-centred approach to design appropriate data science processes. We will do this through hands-on practical where we go through the data science process over an applied case.

The session will provide a brief and will enable you to follow on a scenario where you are expected to act as a data scientist team acting on a task that you will define. We will make references to some human centred design tools to help guide some of our thinking and discussions to guide the process.

35.1 Supportive Reading & Resources

• A good reference from Design Council on their human centred design process where bits of the methodology could frame approaching Data Science projects – https://www.designcouncil.org.uk/news-opinion/what-framework-innovation-design-councils-evolved-double-diamond
• IDEO’s The Field Guide to Human-Centered Design is a broader and detailed guide that can provide some additional methods to approach for the design of DS solutions: https://www.designkit.org/resources/1
• A light practical article titled “Integrating Personas in User-Centered ML Model Development”: https://towardsdatascience.com/integrating-personas-in-user-centered-ml-model-development-afb593741c49

35.2 Further resources

• An interesting project that looks into developing some basic templates to organise data science projects – https://drivendata.github.io/cookiecutter-data-science/
• And a further ethics checklist from the same team – https://deon.drivendata.org/ [Do you see any gaps/issues in their handling of this checklist?]