IV. An Application to Urban Redevelopment

A concept of how an urban design GIS might work to help build a descriptive model of the city has now been presented. The question of what this means to the formation of policy and design interventions has not been addressed intentionally because those questions are beyond the scope of this model. As mentioned in the literature review of environmental modeling, I feel that too often models are constructed as decision making tools instead of decision support tools. The actual decisions must be made by the individual planner or planning agency, and any attempt by the GIS to guide the final decision is risky, as it may lead to a narrowing of possible options too early. This section will look at a current redevelopment proposal for a public housing project and describe what can be learned about the housing development using the model that has just been built.

Redevelopment Objectives

The objectives of the design guidelines are:

  1. Build a minimum of 535 public housing units within a larger mixed income community. These units must be built to a high standard of quality and designed to current standards for family housing.
  2. Reorganize the site to create a typical urban family housing neighborhood, one which cannot be readily recognized as "public "housing", one in which design promotes sustainable safety, and one which is integrated into the urban fabric surrounding it.
  3. To the largest extent possible, this redevelopment must address edge conditions around the site which could compromise or threaten the long term viability of the new community.
  4. Maximize this unique opportunity to coordinate the efforts and resources of different city departments and non-profit agencies including, but not limited to the BHA, MMTTF, BRA, Public Facilities Department ("PFD"), Police Department ("BPD"), Parks and Recreation Department, Community Centers, School Department, Department of Public Works, Transportation Department, Public Library, MBTA, the Mission Church, Medical Academic and Scientific Community Organization ("MASCO"), Wentworth Institute and numerous additional social service providers.

(emphasis added)
Request for Development Proposals[20]

The design guidelines are presented here by the owner, the Boston Housing Authority (BHA), to prospective developers. Although they are for a specific housing project, they reflect the current thinking in how to make public housing work. Major design issues center around supporting mixed incomes, making appearance similar to private housing, and encouraging safety through design (this concept has been put forth by Jane Jacobs as "eyes on the street"[21]).

Site analysis

Design objective one
Here the BHA requests that a large number of public housing units coexist with market rate units. Since public housing is only made available to very low income families, this means that people with incomes that allow them to purchase market rate housing will live next to people who have almost no income. Incomes are detailed in census data, but the lowest level of aggregation is the block group, which covers at least a few city blocks. A unit by unit analysis is therefore not possible with census data. The parcel database affords a better resolution for analysis, but the only data available is land use, land value and building value. An analysis can be attempted by looking at residential building values and finding places where a wide range of values are near each other (Figure 24), but building value is still a poor surrogate for income, which is the stated criteria.

Ideally, the urban design GIS could offer the planner insight into creating mixed communities with the pattern finding application if unit by unit income data were available. One procedure would be to find neighborhoods in which the desired range of incomes were present. The planner could also specify what percentage of the community should fall within certain income sub-ranges. After places that met the criteria were identified, other household characteristics could be studied such as rent, education, place of birth, length of occupancy, building age, parking, distance from public transportation, etc. to further develop an understanding of the composition of a viable mixed income community. As mentioned, this analysis can not be performed well given the current level of detail in census data, but it may be redundant anyway, as the fulfillment of objective two may be the catalyst for the fulfillment of objective one.

Design objective two
The main thrust of this objective is to create a development in which urban form and demographic characteristics blend into the larger community (the safety objective will not be addressed here because safety-related information is not part of the database). Most public housing projects currently follow the design ideology in fashion during their construction in the 1960s and 1970s. They are developments consisting of large, tall, unornamented, multi-family buildings with a great deal of public open space around them. Organic neighborhoods in Boston are townhouses with back alleys that can be used for parking or gardens and façades that come right up to the sidewalk, with a small zone for a stoop and landscaping. Units face streets, which are used by the general public. The streets are narrow, which slows down vehicular traffic and makes the street safer for pedestrians. Most residential buildings are three to four stories tall with none being higher than four and one half stories.

Understanding the difference in urban form is straightforward. Simple visual observation shows a marked difference in building form in plan (Figure 25) and in perspective (Figure 26) between public housing developments and neighborhoods that have formed organically. The planned developments are the antithesis of the organic ones. Residences rarely face the street and public areas relate to the entire development instead of to any one residence. There are few streets in the developments (Figure 27), so no one passes through except people connected to them. South of Mission Main, organic development takes a less dense form as we move away from the city center. There are many neighborhoods of single family homes of one and two stories along with triple-deckers.

Although visual analysis works well here to identify the difference in urban form, an urban design GIS should be able to quantify and operationalize this difference so that the designer gains a another understanding of the difference and so that their methods can be applied over larger areas. The difference in building form might be quantified using an area to perimeter ratio (area / perimeter). The mean of this ratio is twenty-eight for the group of institutional buildings to the northwest of Mission Main, nine for the residential area to the southwest of the development and twelve for the development itself. At first glance, this seems to suggest that a lower building to area ratio would better match the neighborhood, but once again, scale, resolution and the richness of the data set are important because a townhouse can have the same general shape as a public housing tenement, with the major differences being in the architectural detail, building quality, upkeep, number of units and location.

Contrasting the relationship of streets is more straightforward. The public housing developments rarely have streets in them, and when they do, these streets often dead end in the development. Organic neighborhoods contain more streets and they usually culminate in intersections instead of dead ends.

Small nodes seem to be another characteristic of organic neighborhoods, like the BCA node in the South End, but they are absent from the immediate Mission Main area. Two minor nodes can be found south and west of the development, but these are associated with other neighborhoods (Figure 27). If the "home zone" of Mission Main residents could be expanded to include this area, this might not be a concern, but as it is, residents only seem to consider the development as familiar territory.[22]

So three conditions have been identified whose spatial characteristics differ markedly in planned neighborhoods versus organic ones, building form, streets and nodes. Many others probably exist and each designer should be able to use the GIS to develop and quantify their own individual ideas.

Design objective three
Objective three addresses edge conditions. This will be a difficult task as land use data shows that the development is bordered mostly by non-residential uses, except to the south (Figure 28). The development is surrounded by colleges, hospitals, and their associated parking lots. This makes it hard to create a neighborhood environment as there are few existing neighborhoods with which to integrate. Street data shows that the site is also cut off from neighborhoods to the east by a wide transportation corridor.

A designer might begin to look at solving this "wicked" design problem[23] with the tools and methodology developed earlier. Nodes address the problem of the edge between residential and more public areas well, so here is one way in which the edge condition might be addressed. For argument's sake, we begin with the assumption that the designer feels that it is important to create a node in the Mission Main neighborhood. There are presently no nodes in the area because land uses are largely limited to institutional buildings, their associated parking lots and public housing. The designer might explore the possibility of creating a node by inserting commercial, recreational and mixed uses in certain areas of the neighborhood, most likely addressing the edge condition by placing these uses on the northern and western edges of Mission Main (commercial establishments could attract the business of workers in the institutional areas).

By doing this in the GIS and running the node algorithm, a relatively objective judgment regarding the likely success of the design intervention can be generated. With these new land uses in place, the planner can say that given the development of certain areas in the manner described, a certain size node will be created. Other nodes of this size look like place X, place Y and place Z. There are a lot of assumptions in this scenario. Most significantly, the node finding algorithm must be accurate and nodes of certain sizes must be comparable, but at least a design methodology has been created which can be challenged, debated and refined, instead of a design proposal which seems to coalesce in the mind of the designer.

Design objective four
Objective four is to coordinate the activities of all government and private agencies that are working with Mission Main. I included the list of the agencies who would need to be involved to illustrate what a difficult task this would be. In order to accomplish this goal using traditional techniques, meetings would be needed with all of these agencies on a periodic basis. In practice, the coordination of activities is a rare event. This would probably require a great portion of someone from each agency's time every month, and then this person would still have to disseminate the information to the rest of the agency. The knowledge sharing functionality described in the previous section (III. Functional Prototype: Information Sharing) is perfectly suited to address this issue. This functionality was described as it would apply to the situation of a planner organizing input from the general public, but as agencies are only a subset of the general public, the theory is still the same.

The benefits of using GIS to address this objective are numerous. Everyone in each agency would have immediate access to everyone else's latest updates to the database, the need for time and effort-consuming face-to-face meetings would be reduced, and all the information would be accessible on any desktop with a computer and a network connection. In addition to these efficiency issues, sharing information in this way allows the designer to integrate everyone's information into analyses of the kinds described above. Using this system, more information is likely to be considered as part of the project, and the actions of agencies is more likely to coordinated.

Policy and design implications

In order to complete the picture of how these tools might be used, here are a few insights that the above study would support. This cursory site analysis leads to a few definite recommendations. First of all, in order to attract a mixed income clientele, Mission Main buildings should look similar to townhouses in the downtown area. It was mentioned that there is also a precedent for a less dense pattern of single family homes and triple-deckers. Although the argument could be made that the site is outside of the downtown area and therefore more similar to less urban parts of Boston, following this example would make it difficult to accommodate the required number of units on the site.

It will be difficult to integrate the development into the larger community because the site is largely isolated from other residential neighborhoods. Any efforts in this area should focus on connecting Mission Main to areas south of the development, the only direction which offers an opportunity for connection to a larger neighborhood. As the non-residential areas surrounding the development consist of major institutions (hospitals and colleges), there is little that can be done towards addressing edge conditions by changing the physical form of these places. It would make more sense to emphasize that edge and focus energy on connecting the development to southern and western neighborhood areas.

Figure 24: Identifying "Mixed" Communities

Figure 25: Urban Form in Plan View
Free Market vs. Public Housing (arrows point to public housing)

Figure 26: Urban Form In Perspective View
Free Market vs. Public Housing
 (a) 1960s/1970s-era public housing (Orchard Park)

(b) A highly desirable pre-World War II-era high density free market neighborhood (South End)

Figure 27: Nodes around Mission Main
 Figure 28: Land Use around Mission Main

Next section (Conclusion)