Prisons and other carceral facilities are exposed to a variety of environmental hazards such as air pollution, poor water quality, proximity to hazardous waste facilities, and extreme weather events – a pattern referred to as toxic prisons. The proliferation of geospatial technology and modeled datasets has provided data at global scales of coverage, and high spatial and temporal resolutions, enabling diverse stakeholders to leverage measurements of the earth relevant to monitoring toxic prisons. 

Much of the emerging quantitative research in this area perceives the issue of toxic prisons primarily as a data gap needing to be filled and addressed through technological fixes and reforms to prison building. Narratives from formerly incarcerated people, as well as considerations of structural power imbalances, reveal limitations of proposed solutions. In this essay, I argue that while the increased availability of geospatial data for making environmental harm in carceral landscapes more visible to the public is essential, the issue of toxic prisons and other environmental justice (EJ) issues are increasingly at risk of what I refer to here as detached datafication – the uptake of data applied to an issue without analysis of power and without context of those who have lived experiences on this issue, in the data collection, methodology, and presentation of results. 

Reflecting on several examples of data-driven studies and proposed solutions, alongside analysis of power and the lived experiences of directly impacted people, I argue that detached datafication can lead to “solutions” that ignore dimensions of what is needed to fully address toxic prisons. In some cases, they could even create new forms of toxic prisons. In contrast, I suggest restorative/transformative data science as a methodology that can systematically address the inequality, oppression, and violence that produce and sustain toxic prisons, among other environmental and social justice struggles.