Data activism: Occupy Sandy’s canvassing practices after Hurricane Sandy
If activism is about changing the relations, assumptions, and contests pertaining to power, then data activism is about using data as a central tactic to make these changes. “Data activism” is not a term of art academia, but it is used in activism, where it usually refers to persuasively leveraging data to launch action. After Hurricane Sandy, there were many instances of community-based organizations (CBOs) and grassroots responders collecting data, some of which was used to make arguments for certain types of aid or policy decisions. However, rather than looking at how data are used to make political arguments, this post focuses on the practice of door-to-door data collection, or canvasing, as a potential form of activism. That is, I will focus on data collection, rather than using data after the fact, as a site for challenging power dynamics and creating systemic change.
When Hurricane Sandy hit the New York City area in late October 2012, Occupy Sandy (OS) coordinated the largest relief effort in New York City’s history, mobilizing over 60,000 volunteers before Christmas,[i] more than four times the number deployed by the Red Cross.[ii] OS is still active today in many of the communities hit hardest by Sandy. Occupy Sandy grew out of a coalition of activist groups, including Occupy Wall Street, and many of the politics their data strategies employ are legacies of the wider Occupy movement: transparency, community-based governance, horizontal power dynamics, inclusivity, and a drive for justice. Occupy Sandy, like many other first responders, collected data to coordinate disaster relief– to figure out who needed what where, and then organize volunteers and goods to meet those needs. Canvassing
Canvassing involves knocking on doors and talking to survivors to fill out surveys. Regardless of whether the canvassing group is from the local community or the mayor’s office, these interpersonal interactions shape resulting data. In the image above, the National Guard is aiding the NYPD and Global Dirt, contracted by the city, to canvas areas hit by Sandy. These neighborhoods often had traditionally tense relationships with police, especially given protests against NYPD’s Stop and Frisk policies at the time. In general, people are scared and anxious after a disaster. Some people did not answer their doors at all. So fear and trauma, as well as generosity and care, shape the data landscape– they affect what kind of data you can get. RAP & Clipboard Politics Occupy Sandy designed canvassing trainings to address some of these issues. Occupy Sandy trained canvassers in RAP, short for “rapport:”
The slide above is from an OS canvassing training presentation. “RAP is a structured conversation we have with contacts to get us the data we need to establish accountable FOLLOW UP while enabling freedom to CONVERSE with an emphasis in listening.”[iii] Rather than asking residents survey questions in a back-and-forth format, “Have A Conversation: If they want to talk with you about something else, let them. That may be what they need: just to talk. Try to listen between the lines for their other needs: heat, food, medicine and so on.”[iv] They were also taught what one trainer called “clipboard politics.” She explained:
“if you approach someone and the clipboard is blocking you, between you and the person you’re interviewing, that creates a power dynamic. I’m writing things down. You can’t see what I’m writing down. As a canvasser, you slightly change the question. That’s the nature of having a conversation with someone. So they can’t actually see what you’re asking. So it makes them uncomfortable. And in my experience, particularly folks where English isn’t their first language, it’s another layer of– depending on different people’s relationship to being asked questions at their door from the places they come, there’s different reactions. So in Staten Island, the places that were hit, there’s a very large Russian and former Soviet Union immigrant population. Many of them are just like, “I’m not going to answer any of your questions.” For a lot of them, they grew up in Soviet-era governments. You don’t really want to be telling your government that much. So sharing, letting people see the clipboard, asking if they just want to fill it out themselves, if they don’t understand a question, being like, “oh, do you want to read it?” And making it more of a conversation, then that the clipboard is a shared space. That’s what I meant by clipboard politics.”
In effect, through an ethos of “people first, data second,” Occupy Sandy increased the patchiness of their data. Disaster data is always already patchy, incomplete, and messy. Gathering the data is patchy because there is infrastructure missing–electricity, road signs, roads! cell reception, house numbers. People are gone or do not answer their door.[v] Some of these forms aren’t archived or stored. They go missing. They get wet. Post-it notes fall off. The information in the margins don’t fit in the database. Volunteers coordinating canvassing are there one day but aren’t back another. Disaster data is patchy not because of mistakes– patchiness and incompleteness this is the very nature of disaster data, no matter who collects it. However, OS RAP methods actually produced even patchier data.
By having conversations rather than Q&As, by inviting people not to answer questions they were uncomfortable with, and by listening instead of talking, some OS “data stewards”–the name given to volunteers who specialized in data and tech, thought that the survey data became less complete and more patchy:
“I’ve been cleaning [our data]. A lot of the surveys are incomplete, or have one question answered. Probably what happened was that person was with that household for forty-five minutes, right? There was this one survey that had a name, address, contact, and then all these things that they need. And that’s it. It’s a two-page survey. Nothing. Probably what happened was that volunteer who was canvassing was just talking to them and all those questions got asked and he or she just didn’t mark it down or put the clipboard down. But probably that was a beautiful conversation that had to get—you know, not enough answered. So that [survey] got taken out of the sample [and didn’t make it into a data set].”
So much so that many stewards considered their survey knowledge information rather than data:
“That was the difference between what we were doing compared to what something like Global DIRT [who was contracted by the City], which was producing data sets. We have a whole bunch of individual cases that we follow up with individually. And actually, they don’t aggregate.” “The data that we have is so spotty it hasn’t even really occurred to us to examine it kind of holistically and to look at it as a data set. We’re really looking at it as individual cases.”
Like Occupy Wall Street, OS valued transparency, inclusively, and horizontal relations where power was as even as possible. These politics manifested in their data collection practices, and in turn affected the type of data they were able–or unable–to collect. But the methods that put people before data were central to their politics and a model of how they wanted the world to work; it was a microcosm of the “better world” OWS and OS sought to emulate.
The vast majority of information collected by Occupy Sandy’s canvassing efforts was for on-the-ground triage, not after-the-fact forensics. That is, its purpose was to assess needs and match those needs with resources rather than craft a “big picture” of the storm. As such, most questions were about medical, housing, and food needs. Yet, Occupy Sandy forms also asked residents about systemic issues, such as their interest in community organizing, creating and attending direct actions or protests, worker-owned cooperatives, community budgeting, or climate change and justice projects. The table below summarizes some of these questions and, when available, the percentage of respondents who answered in the affirmative.[vi]
At a practical level, such questions are about locating and matching potential community leaders to needs and gauging interest in activities. However, on a political level, the questions expanded the types of crises that Sandy brought to light– precarious employment or unemployment, catastrophic climate change, wealth inequity, and lack of political transparency and participation (particularly at the municipal level)– and sought to imagine possible futures that did not exist in dominant recovery discourses. Most recovery narratives, particularly those used by the City and mass media, advocated for a return to the status quo, while OS survey questions pushed for more economically equitable futures following Sandy.
“I’d say that because of a disaster, unlike most anything else, you actually have an opportunity to try to recover, because […] authentic recovery really means some serious innovation about how communities work and how they derive their power and their resources.” “[By December], we were dealing with poverty, not post-crisis scenarios. It wasn’t about the fact that it was a post-crisis—it was that people have been poor in these areas and will continue to be and so they’re going to come get stuff [from the relief hubs]. And it stopped being about meeting the needs of the storm.”
While worker cooperatives and a participatory budgeting group closely affiliated with Occupy Sandy are now established in the Rockaways, and a strong community movement seeking neighborhood land buyouts has emerged on Staten Island, these concrete “wins” are only part of the activist work the survey questions accomplished. A strong discourse within Occupy Wall Street was the power of imagining different, better futures through smaller scale actions and examples. These survey questions and the way they were asked (via RAP), aimed to foster new political subjects, new discourses, and new concepts of crisis and recovery. In short, they supported a representational politics of possibility—what the future could be like, based on conversations with ordinary people. OS imagined a paradigm of recovery characterized by community autonomy, collectivism, and self-governance. Using Philip Agre’s term, Occupy Sandy introduced “grammars of action” in their canvassing data collection practices: “what matters in each case is not the sequence of ‘inputs’ to ‘outputs’ [into a database], but rather the ways in which human activities have been structured.”[vii] Occupy Sandy’s data collection formalized interaction patterns that aimed to open up possible futures rather than foreclose upon them, to define disaster in terms of ongoing struggles rather than merely as an extreme weather event causing local damage. Special thanks: members of OccupyData, interviewees in Occupy Sandy, Superstorm Research Lab. Funding for this research was provided by the Institute for Public Knowledge, NYU Green Grant, and ISTC-Social. Interviews cited here were collected collaboratively with members of Superstorm Research Lab, a mutual aid research collective. You can get the public data set on OS canvassing here.
[i] Federal Emergency Management Agency (FEMA), Hurricane Sandy: Youthful Energy and Idealism Tackles Real World Disaster Response, Lessons Learned Information Sharing report. Washington, DC: FEMA, August 22, 2013.
[ii] American Red Cross. “Red Cross Recovery Efforts to Help Sandy Survivors,” American Red Cross, first published December 19, 2012, accessed September 2, 2013.
[iii] Occupy Sandy. “A Short Canvass How-To.” Capitalizations in original Power Point slide. n.d.
[iv] Occupy Sandy. “Occupy Sandy Tips on How To Talk With People While You’re Canvassing,” November 9, 2012. 520 Clinton version. Hand out.
[v] Henderson, Tammy L., Maria Sirois, Angela Chia-Chen Chen, Christopher Airriess, David A. Swanson, and David Banks. “After a Disaster: Lessons in Survey Methodology from Hurricane Katrina.” Population Research and Policy Review 28, no. 1 (February 1, 2009): 67–92.
Finch, Christina, Christopher T. Emrich, and Susan L. Cutter. “Disaster Disparities and Differential Recovery in New Orleans.” Population and Environment 31, no. 4 (March 1, 2010): 179–202.
[vi] OccupyDataNYC. “Occupy Sandy Canvass Data.” GitHub, 2013.
[vii]Agre, Philip E. “Surveillance and Capture: Two Models of Privacy.” The Information Society 10, no. 2 (1994): 101–27.