The MIT Press, EAN: 9780262044004, English, 328 pages, 2020, USA

Data Feminism is perfectly framed by an opening gambit that asks: data science by whom and for whom?, arguing that we understand data as “intersecting systems of power and privilege” that “never, ever speak for themselves.” While the overarching approach explores data politics through the lens of intersectional feminism, each of its seven chapters is structured around a core principle. These go far beyond commonly-discussed dataset-induced biases, to systematically address unequal power structures and promote “data justice” as opposed to ‘data ethics’. Some emblematic cases include: the inexplicable lack of data on maternal mortality, particularly black maternal mortality, which by some estimates is three times that of white women; and the ‘Gender Shades’ project, led by Joy Buolamwini (Algorithmic Justice League) and Timnit Gebru, which identified the high likelihood (40 times that of white men) that facial recognition technology will misclassify Black women. This valuable book reinforces the concept that there is no one technology that can fix social problems or indeed fix another technology, but that critical thinking, literacy, and a context awareness are the crucial tools for dealing with the issues surrounding the proliferation of data. It is worth noting that Data Feminism is also available through Open Access.