Practitioner’s Guide
Data Ethics Toolkits v1.0
We currently live in a world full of data and an explosion of effort trying to understand it. In recent years, we’ve seen the increasing usage of algorithmic techniques in order to process raw data and transform it into information. With this comes levels of both risk and uncertainty and the need to explain what these tools are doing, and why. To ensure we build to a fairer and equal future we must ensure we have both fairness and explainability of algorithms.
Within this handbook, we attempt to analyze and critique some key tools, packages and software for fairness and explainability of algorithms, in order to assist you when developing techniques around data.
Our Community
This guide, has been created by the DataKind community on a deep-dive day, where we all worked together to analyze and critique some of these tools. Of course, we by no means believe this is a exhaustive list, and would love to encourage the community to provide us with feedback and suggestions.
- Challenge: There has been an explosion of toolkits, packages, and software for fairness and explainability of algorithms. It is difficult to identify the right one to use.
- Goal of the day: Review the pros and cons of each toolkit and compile a “Practitioner’s Guide to Data Ethics Toolkits”