Open Data has become a prominent ideal in humanitarian information work and is increasingly promoted for crisis situations to increase effectiveness, accountability, and empower citizens. However, like all socio-technical systems, open data platforms for disasters make implicit and explicit assumptions about data, data users, disasters, and the context of use. In this paper, we turn to feminist theory to examine three open data projects rolled out in the aftermath of the 2015 earthquake in Nepal. We used the seven principles of Data Feminism introduced by D'Ignazio and Klein to design an evaluative framework for the three projects. We use this framework to highlight and link the socio-political nature of both disasters and open data platforms. In our results, we highlight significant gaps in how these projects made labor (in)visible, engaged with affective aspects of disaster, addressed context, and challenged power. We argue that these gaps are reflective of dominant practices in open data for disasters and serve as opportunities for designers and crisis informatics researchers to reimagine the potential of such projects. We propose four ways of doing so based on feminist principles and values.

Technology mediated environmental data increasingly shapes how we understand the world, including pressing ecological issues such as disasters and climate change. However, like all data, environmental data is limited and partial. It is necessary to attend to the decisions and practices that create and use this data to understand their limits and advocate for alternatives. In this paper, we draw on postcolonial, decolonial, and anti-colonial theory and a case study of a multi-hazard disaster and climate risk assessment project conducted in Nepal to examine a potential limit of contemporary environmental data practices -the potential to extend or reinforce colonial knowledge systems and extractive relationship to land. Through our analysis, we draw attention to how environmental data practices, such as disaster risk assessment, may contribute to ongoing colonial relationships by privileging technocratic Eurocentric knowledge, conflating disaster effects with economic loss, ignoring ecological impact, and overlooking historical and ongoing power hierarchies. We build on our findings to think through opportunities to reimagine disaster and climate risk beyond probabilistic quantitative models. To do so, we propose four tactics towards an anti-colonial science of risk, as well as argue for a more thorough analysis that attends to situated practices of creating and using environmental data.