Data Study Group Final Report: Accenture

Fairness in algorithmic decision-making

We were tasked with evaluating fairness in its myriad forms, and mapping the various expressions of fairness to the data science workflow. Accenture challenged us to aggregate and organise the elements of the fairness literature into a manageable structure, and to provide meaningful visualisations that facilitate productive discussions around fairness in an analytical project. In this study, we focus on financial services and, in particular, on credit allowance in retail banking, where there is a prevalence of algorithms impacting customers and the services they are able to receive.