Detailed specifications and sources for the underlying data included in this tool can be downloaded here. Keep reading below if you are interested in our overall methodological approach.
The Toxic Prison Mapping Project, Beta, is meant to be a starting point to create a public resource for real-time and long-term monitoring of the intersection of environmental hazards and carceral facilities in the U.S. We hope this tool will continue to evolve with new data and feedback from the public, especially those who are directly impacted by incarceration. Here’s how we developed this version:
Community Engagement Strategy
The project was pursued through a participatory community-engaged approach between academics and grassroots organizers, activists, and other stakeholders working across movements focused on both prisons (inclusive to jails and detention centers) and environmental justice (EJ). This involved three phases. The first phase consisted of 22 one-hour semi-structured interviews conducted by Dr. Ovienmhada between July 2022 and May 2023 with community organizers who work at the intersection of prisons and EJ; about 1/3 of these organizers were also formerly incarcerated. The results of these interviews were published in a peer-reviewed study, Ovienmhada et al. (2023), which characterizes the potential or limitations of data to support community organizer's ongoing advocacy activities related to prisons and EJ. The study findings pointed to specific types of environmental data that organizers needed, as well as cautions and guidelines for how organizers wanted to see geospatial data be employed and towards what ends. Specifically, organizers encouraged geospatial data scientists to develop analyses in ways that support prison abolitionist agendas of decarceration and prison closure.
The second phase consisted of a 90-minute focus group conducted in collaboration with the Campaign to Fight Toxic Prisons (FTP). The focus group participants included four formerly incarcerated activists working at the intersection of mass incarceration and the environment. The goal of the focus group was to specifically solicit feedback on user design and interaction preferences for GIS tools for a realistic set of users. During the focus group, we reviewed both the Environmental Protection Agency’s EJSCREEN Mapping Tool and the Intercept’s Climate and Punishment Map as examples of geospatial web tools made to explore environmental data (the latter examples specific to carceral facilities). We asked the focus group participants for general reactions to each platform as well as specific questions regarding the available data indicators, interaction options, and utility of each tool as it relates to their prison EJ activism. The information obtained from the focus group pointed to several specific environmental data indicators to prioritize as well as query/filter interactions that would be useful to activists.
The third phase involved longer community engagement between MIT and members of FTP in a multi-year participatory technology design process, led by Dr. Ovienmhada. FTP was the primary community partner engaged in the data and software development process through near-weekly meetings from August 2022 to present. Through the three-phase community engagement strategy we developed a list of community-defined desired functionality and data needs for a geospatial tool to support prison EJ organizing. The full list of expressed functionality and data needs are shown in the table below.
Embracing Mixed-Methods: Data Selection and Development
Drawing on the diverse methodological and disciplinary backgrounds of our team, we embraced a mixed-methods approach utilizing quantitative and qualitative ways of knowing from the geosciences, geography, environmental sociology, criminology, and more. Our data on carceral facilities and environmental risk factors come from a wide range of sources - the government, data from other researchers, or data developed from members of the project team. As noted before, the full list of data sources and specifications can be downloaded here. Below are some high-level descriptions:
Quantitative/Geospatial Data
Carceral Facility Data A dataset of 6,738 carceral facility boundaries was obtained from the Department of Homeland Security (DHS). This dataset includes the geographic boundaries and attributes of "secure detention facilities" (in American states and territories) which includes prisons, jails, detention centers, re-entry facilities, jails, juvenile detention centers and more. The Hazards tab includes all of these polygons. The Insights tab includes 1,614 state and federally operated prisons through a semi-manually filtered dataset that can be found on github here.
We also include some descriptive information about facilities including the presence or absence of air conditioning and the age of the building. This information is helpful for understanding factors that may compound with environmental burdens to create high-risk environments. This information was obtained through state-level Freedom of Information Laws by reporters Disha Raychaudhuri and Clare Trainor at the news agency Reuters. The information in the tool represents only a fraction of facilities and will be continually updated at a repository that is not yet public.
Environmental Data The environmental data on the Hazards tab were all sourced from free regularly updated public sources such as the National Weather Service, the National Aeronautics and Space Administration (NASA), and the United States Geological Survey that are all linked on the left 'Data Sources' pop-up on the Hazards page. The code and historical datasets for prison-specific outdoor air temperature, land surface temperature, and particulate matter 2.5 air pollution can be found on github here. The wildfire risk and flood risk data come from models used in the Intercept’s Climate and Punishment map. As described by the Intercept and the original sources, these models have limitations in their accuracy and should not be taken as indicative of true exposure. Future work will aim to map true impact of observed floods and wildfires on prisons. The toxic facilities proximity data, developed by Mothes and Hunt (2023), is a percentile representing average proximity to three different types of toxic facilities regulated by the EPA. The datasets represented in this Beta version do not and cannot fully describe the environmental burdens faced by incarcerated people, however we plan to continually add new environmental indicators in the months and years to come.
Qualitative/Narrative Data
Quantitative and geospatial data are powerful tools for analyzing the scope of environmental injustices incarcerated people face. However, in order to understand and hold space for the depth of impact these toxic conditions have on people’s actual lives our team also incorporated qualitative data in order to learn from, and provide the opportunity to share, the narratives of formerly incarcerated people and their loved ones. This could also be thought of as the "fourth phase" of our community engagement strategy.
The narratives in the Storytelling tab were obtained from semi-structured interviews conducted by team members between August 2023 and August 2024. The methods used to obtain the study data have been reviewed and approved by Massachusetts Institute of Technology's Institutional Review Board (IRB), the Committee on Use of Humans as Experimental Subjects (COUHES) pursuant to Federal regulations, 45 CFR Part 46.101(b)(2), protocol 2210000784.
This research project served as an opportunity to train EJ organizers and activists in FTP, all of whom are outside of academia, in qualitative data collection including conducting interviews, processing and coding of transcripts, and broader approaches to ethnographic field engagement outside the scope of the current project. This training was led by FTP member Jordan Martinez-Mazurek. Dr. Ovienmhada and all of the FTP members of the project team were covered under the IRB process, including obtaining CITI Program certification in Human Research, and could participate in all stages of interview and data processing. The design of this project shifted the nexus of where knowledge production typically occurs, and enabled the grassroots collaborators in FTP and the research team lead by Dr. Ovienmhada to be on equal methodological and decision-making footing as collaborators across the full scope of the project.
Recruitment for the interviews began with FTP’s existing relationships with incarcerated and formerly incarcerated communities, then we snowball sampled from there. Interviewees were compensated $100 via check for participation. We interviewed formerly incarcerated people, as well as people who have loved ones currently incarcerated. Of note is that we did not directly interview currently incarcerated people. The IRB identifies currently incarcerated people as a vulnerable population at special risk of coercion and being taken advantage of. Due to this the IRB requires numerous protections when conducting research with currently incarcerated people that would have presented significant barriers for the scope of this current project. Further, the nature of our project, collecting narratives about toxic prisons, could put currently incarcerated people at additional risk of retaliation from prison officials. For these combined reasons, we chose not to pursue the process to directly interview currently incarcerated people. That being said, some participants in the study elected to receive the list of interview questions in advance and (independent of our study) prepared their eventual responses in communication with currently incarcerated friends and loved ones.
The interviews were difficult for many of our participants who were recounting traumatic experiences that they may have suppressed or in some cases forgotten. We strived to acknowledge this, affirm and validate their experiences, and make space for breaks as needed between particularly heavy moments. Participants retained the right to stop the interview at any time if desired without risk of voiding compensation.
Interview data was processed collaboratively by members of FTP, especially Mei Azaad, and the academics within the team. Another key aspect of our methodology was our approach to processing the interview data. Our research team was very aware of how the carceral state thoroughly strips currently and formerly incarcerated people of their own agency in many dimensions of their lives. Additionally, academic research often strips directly impacted people and communities of the agency to both tell and own their own stories and histories under the guise of anonymity and through the incentives that academics have for knowledge to be attributed primarily to themselves. Due to these dynamics, our team felt it important that our participants had full control and choice of if and how they wanted their interview data incorporated into the Storytelling component of this tool. Each interviewee was provided with consent forms that gave the option to consent to a range of anonymity and format options in the final tool including the choice of whether to use real names or pseudonyms, job titles, direct quotes as text, and/or direct quotes from the original interview recording in their own voices.
We selected quotes to feature on the Beta version of the tool based on connection to specific facilities and to the topic of environmental hazards. Due to the experience of incarcerated people often being bounced around multiple facilities, at times we faced challenges attributing a narrative to one facility. We did our best to omit ambiguous cases. After prototyping the Storytelling tab with these selected quotes and the levels of anonymity selected on participant consent forms, we shared the tool with our participants to give them an opportunity to alter their consent or inclusion in the project. We welcome formerly incarcerated people or people who have loved ones currently incarcerated to contribute their stories to this project by visiting the Contact page. In the future, we also plan to collate narratives previously shared through years of journalism by a variety of reporters and other researchers.
Tool Development Approach
We used the list in the table at the end of the “Community Engagement Strategy” section as a starting point for design, prioritizing the aspects that were most readily achievable with our current resources and capacity at the time. The software design and development was led by the web mapping services company Blue Raster who developed the tool using ESRI products. We want to acknowledge that there is an active effort that has emerged from the Abolition in Geography community to boycott using ESRI products due to their support of predictive policing tools. Our team carefully considered this, centering the perspectives of those on the team that have long participated in grassroots abolitionist organizing. As a team, we decided that using the ESRI environment which provides a robust infrastructure with low-code development was currently the best way to ensure that our tool remains accessible to directly-impacted community members and organizers. We acknowledge this tension in our work and strived to enact numerous other ways for our project to resist carceral logics and support people directly impacted by incarceration. If you have concerns with this approach, we'd be happy to hear your thoughts in the feedback section of our website!)
Dr. Ovienmhada acted as a project manager for the tool development. This work began with full team meetings every other month beginning in Fall 2022 to orient everyone to the task at hand and general scope of the project. In addition to refining requirements through community engagement, Dr. Ovienmhada pursued dataset development, discussed previously.
Entering Fall 2023, the team began to develop wireframe prototypes of the tool, led by Blue Raster. Dr. Ovienmhada worked to translate the community-defined data and functionality needs to Blue Raster and moderate feedback sessions between Blue Raster and FTP from Fall 2023 into Summer 2024. We iterated through approximately 2 - 3 versions of the tool scope and general functionality. This version represents a Beta version for public release that we expect to continue to iterate on in the future to better address more community-defined data and functionality needs.
