@fcorowe)Steif (2021a): “the biggest challenge is not teaching planners how to code data science but how to consider algorithms more broadly in the context of service delivery.”
Steif, K., 2021. Public Policy Analytics: Code and Context for Data Science in Government. CRC Press. A great recent examples of service delivery or data science for social good has been COVID-19. In turn, COVID-19 changed the way that governments think about data science.
Recommended reading: Steif (2021a)
Policy markers (and often decision makers in general) don’t have time or data science training but they have domain expertise. So learn how they make decisions. Learn what the inputs, outputs and consequences of their decisions are.
A set of guiding principles used to achieve a set of goals or outcomes.
It’s key for data scientists to understand policy, to design solutions e.g. time scales, interrelated policies and outcomes and geographic differences.
Programmes are the tool used to implement a policy interventions to achieve specific policy goals.
Good, effective programmes are hard to design. They can fail, even if policy goals are well defined.
Programme evaluation is key but often ignored.
Program evaluation is the art and science of estimating the efficacy of programmes.
We are interested in understanding programmes - often estimating the causal impact of the programmes on an outcome of interest.
Steif (2021a): “This is operational decision-making, and the area where data scientists can have the greatest [immediate] impact. The goal is to deliver resources when, where, and to whom they are most needed.”
Solve problems
Research is as a major activity in government and private corporations in more developed countries
The research problem statement is the most important part of a project. It must be fully and clearly phrased to represent the goal of your research. It should be written in a complete grammatical sentence in as few words as possible.