For four hours, one voice continued to be documented. Jen was relentless, demanding more trees, a greater “sense of magic”, and a series of specific paths in her neighborhood park in northern Brooklyn. She had to see evergreen trees on the north side of the park, as that was where people sat in the wintertime to catch the sunset: winter screams evergreens. Tasked with soliciting public input on the park redesign, I dutifully recorded her remarks largely without realizing that she had monopolized my focus at the expense of gathering the thoughts and desires of others. Jen’s tenacity, seemingly unlimited disposable time in the middle of a weekday, and spatially articulate vocabulary resulted in her having a greater say in the communication of a public’s preferences for park design to the professional design team. Ironically, when we returned two weeks later with a schematic design for further public comment, Jen was furious, arguing that no part of the scheme reflected her wishes. She cornered me at the picnic table and promptly re-iterated her thoughts on the future of her park.

This experience of managing public participation in north Brooklyn illustrated to me some of the challenges of design outreach sessions. I was especially surprised by the difficulty of understanding the extent of community agreement on spatial opinions. Often, the beliefs of the loudest and most design-articulate people get weighed more heavily in the outreach process as opposed to a majority of possible park users. Today, most outreach events neglect utilizing technology to fully expand analog methods of data collection, while those that do often limit their efforts to GoogleForms for collection and Excel tables for interpretation. Most feedback is considered subjectively, and very little of it is interpreted spatially. I am interested in the potential for computational tools to involve a client/public more directly in the design process and offer the designer themselves spatial feedback points to build from. At the intersection of weird web tools, climate-sensitive design, and the funky behaviors of people online, is there a way that technology can facilitate a spatially-focused discourse? If computational tools can create a visual, data-driven design process, would we collaborate more deeply over public spaces in our contemporary cities?

Prototyping

To explore these questions, I developed a prototype of a user-generated design tool and applied it to my friend Louis’s yard in Bushwick, Brooklyn. The tool uses a web-based form that asks users how important five features are to them on the specific site in question:

• Paved Space
• Lawn
• Habitat Creation
• Trees
• A Pool

Before users get to vote, a model of the site is developed in Rhino, from which Grasshopper runs a climatic analysis to propose the ideal locations for the various features. For example, Grasshopper determines where to site trees by pulling from a list of points on the site that have at least eight daily hours of sunshine. Alternatively, paved space is clustered near points that fall below this sun exposure threshold. Once these procedural rules have been set up, the web form goes live and asks users how deeply they value individual park features. Each response is given a numerical value, which is then averaged per question to result in a percentage of site to be filled by the feature in question. If Louis inputs that lawn space was “Very Important” to him, that value would be recorded as “75” and the algorithm would randomly fill 75% of the space closest to the ideal locations for seeding lawn. If another voter inputted that the lawn was “Not at all important”, that value would be recorded as “0”, the two values would be averaged, and the algorithm would randomly fill the optimal 37.5% of the site with the lawn. As users vote, designs reflecting the overall balance of preferences of the features are continuously visualized in Rhino. Ultimately, the web form includes the option to end the iterative voting should a majority of users vote that they are satisfied with the design. Beyond terminating the cycle, this triggers a “bake” command in Grasshopper that would become a visually-agreed upon basis of design for further refinement.

GIF showing a range of design options voted upon. Note the clustering of trees and lawn in select points on the site.

Louis, Eliza, and Rosa came over for brunch knowing that they would be testing my “grad school prototype” but aware of little more. I distributed the link to the webform and began recording them as they huddled around my computer monitor and voted away. Immediately, they were curious about how the cluster points were generated, with Louis becoming obsessed with checking the sunniest points in the model against photographs he had of his yard (needless to say, the sun model was accurate). Eliza led a charge to turn the whole yard into lawn, then pushed to eliminate all trees and turn it into a sun-tanning plane (needless to say, not everyone agreed). I watched as the invested parties expressed and resolved disputes, thumbs quickly filling radio buttons, and hit submit with excitement or disgust directed at their peers. Eventually, Louis negotiated an end to the iterations by getting buy-in on a specific design and voting to terminate and bake the scheme. I then manually rendered a version of the baked design with locally appropriate tree, sedge, and shrub species, a running bond paver pattern, and some scale figures. I followed up by sharing the design with them for comment.

Slider showing Grasshopper result of collaborative voting (left) and rendered schematic vision (right)

Who is crunching the numbers?
Can we meet him?

An initial concern raised during the testing session of my prototype regarded how the algorithm was distributing participant preferences. The user-driven design tool could be broadly categorized as an automated decision system (ADS) where “decisions on spatial configurations rely on the usage of computerized components to make or influence a decision”¹. Of course, by nature of being a visualization, the tool represents a very low impact ADS whose “decisions” still require a human confirmation before affecting real lives. Regardless, the system certainly featured elements of bias, such as the desire to site trees in the sunniest locations, that might conflict with desired uses of the park. Eliza’s crusade to eliminate shade and open the site up as a sun-tanning haven was inherently challenging given how the algorithm calculated preferences; if solely one user submitted that trees were very important to them, it would be mathematically difficult to completely eliminate that submission in an average of the overall inputs. Perhaps key to combating this bias is the explainability of the ADS: scholars have written on the fundamental importance of algorithm explainability in holding decision making systems accountable, with some analyzing the existence of this concept of required explainability as a legal right in the European Union². Others have observed how automated systems intended to better reflect democratic wishes, such as ranked choice voting, can result in depressed turnout due to confusion over the ADS³. Shifting this scholarly work into praxis, databases have been published that empower individuals to investigate, interpret, and explain important algorithms in daily use by American government agencies⁴. Of course, in the case of my prototype, I simply answered questions on the spot of how the algorithm was working in response to Eliza’s desire for a full sun yard. In order to be sustainably and ethically scalable, however, my computational design tool might require a supplemental explainer page articulating how user-inputted data would be used to generate spatial concepts.

An additional source of disagreement resulted from a review of the weight of voting importance between my participants. Louis felt that as the site in question was his yard, his vote should weigh more. Eliza was frustrated once she figured out that each user could vote endlessly and have these values recorded in the general pool for averaging, resulting in whoever had the fastest thumbs getting more say. Rosa didn’t really care about the whole thing and just tried to create a pool. In some ways, the procedural clustering of the feature zones based on climatic data resulted in greater trust as it tied design options to objective data. However, it fell short of creating a communal sense that the visualization reflected a compromise of the desires of the group, leaving me to mediate and attempt to unify consensus. This navigation of stakeholders in conjunction with a data-driven design process mirrors the conclusions of scholars analyzing master plans that require client buy-in⁵. It also underscored the persistence, even in a radical democratic system, of broader issues with collaborative decision-making at the scale of shared public space. If Louis didn’t pay rent on his yard, would he still feel entitled to a larger say than others due to some other social, economic, or political factor? Would he feel that way due to his geographic proximity, a warm memory located there, or perhaps a connection to a nearby restaurant?

From Historical Precedents to Contemporary Applications

The intersection of democratic representation and public space defines the present and future implications of my computational design tool. Struggles for democratic control over public space has a storied history in the City of New York, with documentation of modern contests over shared space dating back to the initial European contact. In the mid-twentieth century, the explosion of urban renewal projects amidst the context of ongoing socio-racial prejudice, segregation, and violence resulted in radical groups reclaiming and reimagining public spaces. One such group was the Young Lords, a Nuyorican organization that organized around improving conditions for Puerto Rican migrants in the mid-century city. Through protests such as a 1969 garbage fire initiative that criticized a lack of city sanitation services in East Harlem and the Lower East Side or the 1970 occupation of the city-run Lincoln Hospital⁶, the group demanded better conditions in underserved neighborhoods. The actions of the Young Lords and their peers could be read as a form of participatory voting based on spatial occupation of public assets (the street and the hospital, in this case).

From left to right: Image captures a Young Lords protest in El Barrio, July 1969; Protesting the construciton of the world's tallest jail in Chinatown, Manhattan, January 2022; Debate at a contemporary community board in Lower Manhattan, November 2013.

Public space was therefore the site of intense struggle through social movements as well as political movements. The post-urban renewal city grappled with political resistance to the top-down planning of the built environment exemplified by master builders like Robert Moses. This resistance would lead to the devolvement of some planning powers to local actors, which would ultimately take the form of community boards⁷. Initially conceived of as an experiment in participatory planning, the city’s Community Boards offer a forum for discussion of planning projects, local built-environment regulations, and other shared concerns. In some cases, community boards have taken advantage of a specific clause in their charter to propose detailed spatial plans for redevelopment of parcels within their extents, such as the 1992 Nos Quedamos plan for redevelopment in Melrose⁸. Generally, however, the role of these participatory forums is largely advisory and does not allow for approval or disapproval of specific development projects within the community board’s geographic bounds. Some academic work, however, has underlined the potential for the boards to harness the legal powers of urban renewal–eminent domain and blight condemnation, among other powers–to empower participatory forums to execute development projects⁹. In reality, given limited budgets and knowledge of the requirements of master planning and architectural drawing, it is often impossible for community boards to develop the level of detailed documentation necessary for implementation of a development project.

Is there ultimately a through-line that connects radical spatial reclamation and participatory planning systems to… a quirky digital design tool? Could the computational tool, which provides detailed plans that were agreed upon through voting in a bottom-up manner, provide a documented design ready to be fully executed? One possible application of this could be as a response to the contemporary moves to privatize and reuse buildings and lots of the New York City Housing Authority (NYCHA). The buildings of NYCHA form bastions of affordability in the current market of the city–certainly complicated, underserved, and often viciously underfunded structures, but resources nonetheless. The recent move to privatize certain developments¹⁰ and sell parking lots and other open spaces to developers will likely not result in development that represents the wishes of local residents of the housing complexes¹¹. In Edgemere, Queens, where the NYCHA Edgemere Houses were recently rebranded as the Ocean Bay Apartments, the nearby Beach 41st Street Houses have developed a patchwork program for its various outdoor spaces. Given the stated goal to develop the “semi-utilized” lots of its recently privatized neighbor, there is reason to be concerned for the future of the complex’s community gardens lots, informal outdoor spaces, and bay frontage. Could a computationally-informed, voter-verified design command attention in a climate of privatization of public assets and recenter the interests of local residents?

Images of the Beach 41st Street Houses, including the existing garden plots, shade structures, and playground mobile library uses.

A computationally generated for open space in the complex along Beach Channel Drive. Left, the Grasshopper output and right, diagrammatic scheme.

Imaginative Futures

In Speculative Everything, Anthony Dunne and Fiona Raby argue against the contemporary trend for technology to be exclusively bought and sold. They challenge “the idea that something is not ‘real’, when real means it is available in shops”¹², instead proposing a conception of technology as a tool for speculative exploration of human possibility. In many ways, the visualization of potential spatial designs in response to voter-inputs is this speculative fiction, an idealized look at what could be. It excludes the arduous path from public desire through construction, budgeting, and political limitations to public park. In the future, the question of how democracy, speculative as well as real computational tools, and spatial creation engage with tangible spaces will dominate our cities and lives. Perhaps in 2042 there will be computationally-informed, voter-driven design tools that fully supplant designers and result in a radical republic of park rabble. Perhaps a demagogue will utilize such unified, visually-confirmable consent to enact autocratic visions of grandeur. We can imagine what could have been had Robert Moses gotten people to vote on the dramatic renderings of his proposal for the Battery Bridge¹³, or the architecturally fascinating vision for the Lower Manhattan Expressway (LOMEX)¹⁴. There are, no doubt, hundreds of reviled proposals that may have benefited from a computationally-informed decision process.

For me, however, I dream of a future where democratically-accessible computational design tools become the tool of their own invisibility. They would provide a nimble armature for the expression of dissent, agreement, and simple infatuation with public space proposals through excellent visuals and easily explainable voting processes. Next, the tools would work to develop a strong sense of consensus, iterating through options until all parties are uniquely unsatisfied. After such remarkable compromise is brokered however, technology would retreat, lacking a means for its own enforcement, and remain in the background until invoked. In this world of Rule by Humans, should a community wish to fly in the face of the ideal climatic locations for trees, they should be able to. The role of the designer, therefore, may remain uncertain and ever-changing, but the purpose of the designer’s tool remains one of empowerment, conflict resolution, and imaginative thinking without rotating into sociopolitical control. Perhaps in 2042, therefore, computation will not have enabled the rise of a philosopher king, the dictatorial rule of the majority, or sterile proclamations informed by tabulating an extensive CSV file. Rather, it may simply help us achieve that warm feeling of walking out of a room feeling as if you’ve been heard.

References

Younge, Yusuf, Voegeli, and Truong. (2019, April), Automating NYC. https://automating.nyc/#aboutUs.
Doshi-Velez, F., Kortz, M., Budish, R., Bavitz, C., Gershman, S., O'Brien, D., ... & Wood, A. (2017). Accountability of AI under the law: The role of explanation. arXiv preprint arXiv:1711.01134.
Northwestern University Computational Journalism Lab. (2017, January), Algorithm Tips - Resources and leads for investigating algorithms in society. http://algorithmtips.org/.
Donovan, T., Tolbert, C. and Gracey, K. (2019), Self-Reported Understanding of Ranked-Choice Voting. Social Science Quarterly, 100: 1768-1776. https://doi.org/10.1111/ssqu.12651
Amato, Rebecca. (2018, March 08), Board to Death, Urban Omnibus. https://urbanomnibus.net/2018/03/board-to-death/.
Wilson, L., Danforth, J., Davila, C.C., & Harvey, D. (2019). How to Generate a Thousand Master Plans: A Framework for Computational Urban Design.
Fernández, J. (2009). 3. The Young Lords and the Postwar City: Notes on the Geographical and Structural Reconfigurations of Contemporary Urban Life. In K. Kusmer & J. Trotter (Ed.), African American Urban History since World War II (pp. 60-82). Chicago: University of Chicago Press. https://doi.org/10.7208/9780226465128-006.
Soomro, Alia. (2019). Dismantling Mid-Century Urban Renewal: A Community-Based Approach for the Future of New York City, 84 Brook. L. Rev.
Honan, Katie. (2017, January 12), Changes Promised at Far Rockaway NYCHA Houses with Private Partnership, DNAInfo. https://www.dnainfo.com/new-york/20170112/edgemere/ocean-bay-edgemere-apartments-nycha-lease-public-private-partnership/.
Vielkind, Jimmy. (2018, December 12), NYCHA to Open Land to Developers to Raise $3 Billion for Fixes, Wall Street Journal. https://www.wsj.com/articles/nycha-to-sell-land-to-developers-to-raise-3-billion-for-fixes-11544652932.
Rothman, Lily. (2017, April 21), Manhattan Almost Had a Highway Running Across It. Here's What It Would Have Looked Like, Time Magazine. https://time.com/4742520/jane-jacobs-robert-moses-drawings/.
New York Preservation Project. (2011), Brooklyn Battery Bridge, NYPAP. https://www.nypap.org/preservation-history/brooklyn-battery-bridge/.