Author: Kirsten Hoogenakker
Publisher: DataDrive
Publication Year: 2021
Summary: The following article discusses how math has generally been thought of as a completely unbiased form of research. Numbers are numbers, how could they be biased? However, when data scientists do not employ the proper ethical constraints to our work, this perception of a total lack of bias in data can be wildly incorrect. In order to ensure that data and models are as unbiased as possible, there are 4 main tenets of ethical data science to keep in mind: 1). Understanding privacy, 2). Ensuring your training data set is unbiased, 3). Developing transparency in your methods, and 4). Zeroing in on blind spots.