ISCN Global Mixer: Data feminism and its relevance for Smart Cities

Data feminism addresses gender inequalities and recognizes discriminatory power structures in the field of data science. For a conscious handling of data in smart cities, various principles of data feminism can therefore be helpful in detecting and preventing discriminatory patterns.

In this session of the ISCN Global Mixer, Vanessa Hochwald from GIZ Data Lab spoke about Data Feminism and its relevance to Smart Cities. After giving a short introduction about the work and mandate of her organizational unit, which she dubbed as a think-and-do tank, she established two working hypotheses for the data feminist lens of the GIZ Data Lab. These are, drawing from the book "Data Feminism" by Catherine d'Ignazio and Lauren F. Klein, a) Data is Power, b) Feminism is about Power and challenging it in assumption of structural inequalities.

Conceptually, this translates into several principles and operationalizations for project work. In the context of development cooperation, it warrants for example the introduction of design thinking, the setup of gender budgeting or focussing on marginalized groups to build data literacy.

Vanessa then led through some key interventions and examples as to how the data feminist perspective applies to smart cities. The first by now sadly "classical" and notorious examples revolved around the misuse of CCTV images for racial profiling or biases woven into rigid and unfair criteria for the allocation of welfare benefits. Consequences of that according to data feminism should be the application of data minimization and the right to data refusal on behalf of individuals in order to protect them from negative consequences.

Further, data can be used for better advocacy for marginalized groups. The keynote mentioned here a historical example where communities in Detroit were empowered to track where commuters ran over black children, ushering improvements in road safety, or AI models generate visuals for proposals of more feminist and inclusive urban sites.

Another area of intervention can be in the joint design of questions and interventions with the target group. Vanessa's reference here was an algorithm suggesting where the best locations for childcare facilities in Mexico City would be. This is closely related to projects that aim for closing Gender Data Gaps in order to reduce underrepresentations or negative overrepresentations of marginalized groups in datasets for policy analysis and intervention. The Transformative Urban Mobility Initiative (TUMI) for example - which also presented their Mobility Data Hub in one of our past Global Mixers - has done the first large-scale comparative data collection to understand specifically mobility patterns of women in four African cities. 

For a full dive into the keynote, have a look for yourself in the recording above.

Further links:

• Blog of the GIZ Data Lab
• (German) article by executive board member of GIZ Anna Sohpie Herken on Data diversity and inclusive AI [paywall]