Online Tools for Public Participation

Dr. Kheir Al-Kodmany, Assistant Professor, University of Illinois at Chicago

Email, Kheir@uic.edu Phone (W), 312-413-3884 Phone (H), 708-233-1505 Fax (W), 312-413-2314

Address: CUPPA Hall 412 South Peoria Street Chicago, IL 60607

Key Words: Community Planning and Design, Citizen Participation, Urban Likability, Internet Applications

Abstract

As the Internet continues to reshape how the public communicates, planners have the opportunity to use Web-based technologies to widen access to decision making in community planning and design. This paper describes an interactive web-based survey tool that was used to advance a community planning process in the Chicago area. The project involved collaboration between researchers at the University of Illinois at Chicago and leaders and residents of Pilsen, a nearby community. The team used a process taken from the work of Jack Nasar (The Evaluative Image of the City, 1998) who developed a method of surveying residents to determine which areas they like and dislike in their community, with the goal of creating a single "evaluative image" of the community that could guide future design and development. Researchers employed a Web-based interactive map that was linked through a server to a GIS program, rather than the traditional phone survey used by Nasar. This project builds upon an earlier version of the survey in which the map was limited to only one section of the community. In increasing the complexity and coverage area of the survey, we were able to develop a more comprehensive understanding of community residents' evaluative image of the whole community. The paper ends with a discussion of the obstacles and future implications of using the WWW to further public participation in planning.

Introduction

American universities in urban settings are increasingly using teaching, research and service to address quality of life issues in their communities. Some, like the University of Illinois at Chicago (UIC), are beginning to focus on doing participatory planning and design in the neighborhoods where they are located. Over the course of the past three years, UIC researchers and community leaders in the Pilsen neighborhood of Chicago have been engaged in a community planning process that has utilized a wide variety of high and low tech participation and visualization tools, including community design workshops, a community geographic information system (GIS), artist's sketches and Web visualizations (Al-Kodmany, 2000). In the most recent project, researchers have explored using the WWW as a medium for public participation by using new Internet-based map and survey tools to collect local spatial knowledge from the public. The planning team created a Web-based survey that was linked through a server to a GIS.

The theoretical foundations for this Web-survey can be found in the work of Kevin Lynch (1960) and Jack Nasar (1998), who argue for the importance of discovering how city design affects citizens. By studying what kind of evaluative image residents have of their community, planners, researchers and community leaders can derive valuable information about how to improve the physical form of their communities. Nasar developed and implemented a method of surveying residents (using traditional phone surveys and manual map-making) to determine which areas they like and dislike in their community, with the goal of creating a single "evaluative image" of the community that could guide future design and development.

The UIC team advanced Nasar's original method by using the Internet and a GIS program to survey and map residents' preferences. While it has become common for local governments, community planners and community agencies to utilize the Web to offer information to community members, we have taken the next step by using the Internet as a medium for two-way communication; the tool allows people to become both receivers and providers of information. Until recently, the use of spatial data with Web technology has largely been limited to the provision of information to the public through 'image-maps' rather than any actual public participation. The current project is one of only a few projects that involve the public returning spatial data as well as consuming it. The project advances the prospect of allowing citizens to comment, using Web-based maps, upon government actions and development proposals. It is only with the recent development of sophisticated client-server, Web-optimized languages and implementations such as Java / JavaScript / JSP / Servlets and Active X / VBScript / ASP that the field has developed (Evans, et al, 1999).

The paper is divided into 4 sections. Section one explains Lynch's concept of imageability and Nasar's concept of likability and how this can be measured through a survey. Section two provides the background of the university and the community and describes the project, and section three discusses possible ways to improve and expand the technology used in the project. Section four discusses conclusions.

I. The Concepts of Imageability and Evaluative Image

The basic theory underlying our work is Kevin Lynch's concept of imageability. In his seminal book, The Image of the City (1960), Lynch uses the concept of "imageability" as a theoretical framework for studying cognitive maps, urban form, and the spatial relationships of cities. Imageability "is that quality in a physical object which gives it a high probability of evoking a strong image in any given observer. It is that shape, color, or arrangement that facilitates the making of vividly identified, powerfully structured, highly useful mental images of the environment. It might also be called legibility" (Lynch, 1960, 9). A highly imageable or legible city would seem well formed, distinct, and remarkable. An imageable city helps residents and visitors to better spatially orient themselves, to navigate through the city and to find their way. He also explored how regular citizens use and visualize city spaces. Lynch maintained that his theories and models could apply to various scales of spaces, from the largest cities to the smallest communities and neighborhoods.

In The Evaluative Image of the City, Nasar extends Lynch's work. He argues, however, that knowledge about imageability is not enough. His work focuses on meaning, which refers to inferences about the quality and character of the place and its users. Human feelings and meanings define what Nasar calls "the evaluative image of the city." This phrase refers to how the public evaluates the cityscape and how they react to various environments. We recall places about which we have strong feelings, and we will more likely have feelings about the recalled imageable parts of the city (Rapoport, 1970). Research has found that the most imageable buildings in a city elicit the strongest evaluations both positive and negative (Appleyard, 1976). If most people like the imageable elements, the city will probably convey a positive evaluative image. If they dislike them, the city will convey a negative evaluative image, suggesting a need for changes in the city's appearance. This aspect of city image is what Nasar calls the likability of the cityscape. Likability refers to the probability that an environment will evoke a positive evaluative response among the groups of people experiencing it. Inhabitants of a city with a good evaluative image find pleasure in the appearance of its memorable and visible parts.

Nasar suggests that it is possible to learn the public's preferences by empirically measuring them. Just as we weigh objects to find how light or heavy they are, Nasar says that we can measure preferences to determine the degree to which people like or dislike various areas of a city. Nasar employed the evaluative method in two U.S. cities, Knoxville and Chattanooga, Tennessee. His team interviewed 160 residents and 120 visitors. The resident interviews were conducted by phone and the visitor interviews were conducted in person with a city map as an aid. The participants were asked to identify up to five areas that they liked visually and five areas they disliked visually. Interviewers probed to discover the boundaries of the areas mentioned. Then they asked the respondents to state the reasons for their responses. A single map was created for each respondent, representing that person's evaluative image of the city. The individual maps were then overlaid to create two composite maps, one for residents and one for visitors. These maps showed the location and likability of visual features, and the degree of consensus for the evaluation. The maps provided a basis for a visual plan to guide the future appearance of these cities.

II. The Map-Survey Project

1) The Pilsen community and the University of Illinois at Chicago In 1998, UIC planners and designers were invited to be part of a participatory community planning process in Pilsen, which is a primarily Mexican-American community of 50,000 residents just south of the UIC campus. Like many other low-income neighborhoods, Pilsen is faced with a host of urban challenges, including heavy traffic, dilapidated and vacant buildings, and crime. Community leaders felt that the meaningful involvement of a broad range of residents in a planning process would strengthen the overall sense of community, and a cooperative effort would help present a "unified front" when funding opportunities arose. A planning team was formed that included 25 community members and five university professionals: two architects, two planners, and a graphic artist.

At the heart of the UIC-Pilsen planning process was a series of community design workshops where a variety of interactive high- and low-tech tools were used. For the next stage in the process, the research team sought to use new interactive technologies being developed on the WWW to see if the Internet could enhance existing public participation efforts. In particular, we began developing a Web survey, using Nasar's basic framework, to discover Pilsen residents' spatial perceptions of their community. The team intended to employ the Web survey to supplement, rather than to replace, the ongoing face-to-face workshops.

2) Overview Researchers created a Web map-survey tool that enabled two-way interaction with residents over the Internet. Two versions of the survey have been implemented. In the first version of the survey, only a limited area of the community was available for selection. The second version (the project described in this paper), had the same goal of discovering community residents' evaluative image of the community, and similar instructions for selecting areas of like and dislike, but the map was expanded to include the entire community area and a new way to navigate the larger spatial area.

The Web-based interactive map tools have been developed to work on average computers, and do not require the loading of geographic information systems (GIS) or Internet map-server (IMS) applications. At the present time, a GIS program is used only at a later stage for analyzing the composite participants' data and integrating it with other GIS layers. Since there is no interaction between the GIS database and the users (local residents), the users can navigate the maps and photographs without getting bogged down in the complex manipulations of a sophisticated GIS. Local residents simply interact with Web-maps by clicking map locations and typing text to give us their opinions.

Two community-based organizations, 18th Street Development Corporation and The Resurrection Project, facilitated the survey process. They advertised the survey and sent out letters explaining the purpose of the survey and encouraged residents to conduct it on-line. The letters indicated the URL and the password to access the site and invited residents to come to their offices to do the survey if they did not have other access. Also, residents had the option to do the survey at the local library or a university computer lab. UIC researchers and leaders of the community organizations worked together to create as "random" a sampling as possible. Members of the community and graduate student interns did a great deal of work in designing and testing the survey, informing local residents, contacting leaders and talking to business owners.

The goal of the survey was to discover how residents of Pilsen evaluate the appearance of their community by asking what particular places they liked and disliked. The survey instructions emphasized the issue of community appearance and encouraged respondents to think of which places they found to be pleasant and which places they found to be blighted and unattractive. Once all responses were received, we created two composite GIS maps, one showing the areas of the community that were liked and the other showing areas of the community that were disliked. The goal is to use these maps as guidelines for decisions on community appearance and visual form.

3) Areas of study and spatial units In the first project the area of concern was a small commercial strip of .5 mile by .125 mile or 40 acres. The area was broken into 16 cells or spatial units, each of which was .06 mile by .06 mile or 2.5 acres. Thus, the final selection area was about a quarter city block. In the second project the area of concern was the entire community and some of its adjacent areas. The coverage area was 2 miles by 2 miles or 2,560 acres. The area was broken into 256 cells or spatial units, each of which was .125 mile by .125 mile, or 10 acres. Thus, the final selection area was about one city block. The total area of concern in the second project was 64 times greater than in the first project. One result of the larger coverage area was that the final spatial unit that residents could select as a "like" or "disliked" area was four times larger (10 acres vs. 2.5 acres) in the second project.

4) The map-survey interfaces

In the first map-survey design (see http://g015.cuppa.uic.edu/gridfeedbak/xarial18St/arial_18St.html), researchers included only the main commercial area of the community. The survey map consisted of an aerial photo of the selected area, with a white grid overlaid on top, which the user could turn on or off. The instructions directed the participants to consider areas of the community that they liked and areas that they disliked. Before selecting a location on the map, the user was asked to first click on one of two radio buttons labeled "like" and "dislike". When the appropriate button was selected, the user would then click on a location on the map. The square of the grid containing the selection would then be colored in green if "like" had been selected or red if "dislike" had been selected. When a square was clicked, a small text box opened, asking the participant to state a reason for liking or disliking that area. When the user was finished, he/she clicked the "submit" button and the data was then transferred to the server at UIC.

After analyzing results from the first survey, it became clear that in order to be more useful, the survey needed to cover the entire community area, rather than just the main commercial strip. Expanding the coverage area, however, was not a simple matter, as it posed problems in terms of image and file size, resolution and ease of navigation. The original map-survey interface took up a large amount of space on the screen, since the aerial photo was published at the highest possible resolution, six inches per pixel. Also, it required a lengthy amount of time to load onto the user's screen because the aerial photo was set as one image and was not broken into smaller pieces, or tiles, to make the loading faster. Users had to scroll up and down and right and left to see the map and they could not zoom in on any part of it. In redesigning the map-survey, we had to address the issue of scale.

In the new project (http://g015.cuppa.uic.edu/gridFeedbak/final/index.html), we developed a second map-survey interface where we have "nested" three different scales, or resolutions, within one map. The user must "drill down" to reach the final map where they select the location that they like or dislike. The first map on the opening web page is a graphic representation of a paper map of Pilsen with a grid overlaid on top. The second map is a course resolution aerial photograph that allows the display of a greater surface area of the community, but not enough detail for the user to identify particular streets and buildings. The final map is the original digital map at its highest resolution so that users can read the fine details and distinguish buildings and streets. On this final map, the user selects the location that he/she likes or dislikes and then types his/her reasons for the choice.

5) Mapping the survey results Because the Web server was linked to an Oracle database and a GIS application, we had the capability of taking all the points that were selected by the participants, sorting them by longitude and latitude, and plotting them on a map automatically. The Oracle database could also group the associated comments. In this manner, a community-input database was created that contains the range of views about areas liked and disliked with the associated reasons.

Several GIS maps were then created from the survey data, including two "evaluative image" maps (urban likability and urban dislikability). We utilized a choroplethic map to illustrate the intensity of likes and dislikes for locations in the community. Dots were used to represent intensity: the number of dots in each cell of the map was proportionate to the number of times that area was selected by the residents in the survey exercise. That is, areas that were chosen most often, indicating a greater likability or dislikablity, were represented by a greater number of dots and hence a darker appearance. The first map has dots of various intensity showing the areas most liked, and the second shows the areas most disliked. These maps give us the evaluative image of the community, including the well-known (imageable) places that observers selected most often. In addition, these GIS maps are interactive; clicking on an area (or cell) of the map opens a window of text that lists the residents' stated reasons for liking or disliking that area. One idea we want to implement in the future is to ask residents to tell us where they live and then see if there is a correlation between their home location and their choices and comments on the map.

6) Discussion In analyzing the recorded comments, it became clear right away that the comments from the first survey were much more specific than in the 2nd survey. Part of the reason for this is that in the second survey, the smallest selection cell was 1/8 mile by 1/8 mile (.125 mile by .125 mile) while in the first survey the smallest cell was 1/16 mile by 1/16 mile (.06 mile by .06 mile). Also, the area covered in the first survey was a major commercial corridor that was very familiar to all residents; it was easy to relate to and to provide comments. However, in the second survey there were some areas, particularly the industrial districts, which were rarely chosen by residents, so we did not expect to receive very detailed information. Indeed, comments about these areas were very generic. In the first survey people said things like, "I don't like how there is always so much litter on this corner," or "I like how these houses look so new and clean." They also talked about the need for more unified landscaping, more unified signs, fixing and widening sidewalks and the need for more of a cultural touch in the design of facades.

In the second survey, some respondents were not sure what to talk about and the answers were more general and imprecise, such as, "This industrial area is really ugly," "we should have these industrial areas more controlled," "we should have more city regulations of these areas," "the Chicago River area needs to be cleaned," "the river needs to be put to use by local residents, it could be developed as tourist attraction, the present use of it is dumb for the industrial activities." While these answers were less helpful for pinpointing specific planning improvements in the community, they did give us insight into residents' overall evaluative image of their whole community. The composite maps that were created show that large parts of the community were fairly unfamiliar to the majority of respondents. The well-traveled areas received the highest number of responses and certain areas, such as 18th street, the major commercial corridor, were classified as the most imageable parts of the community.

The new survey project that included the larger coverage area had both advantages and disadvantages. The major advantage of the new survey was that we could begin to understand the public's evaluative image of the whole community area, rather than of the limited area of the main commercial district. Yet it was a challenge to apply the original tool to the larger geographic region. In the first survey design, we were able to provide street level visualization aids such as pictures and movies. These helped participants to recall these places. In the new project we could not provide visualization aids because of labor, cost, and technical problems. The size of the geographic area would require a tremendous number of images and would involve many hours of labor to collect the images and prepare them for web visualization. Also, the large size of the files would make them difficult to access and load from the web. Another issue with the larger coverage area was the cost of the aerial maps. We purchased the set of four tiles (each is one mile by one mile) for $500 from an outside vendor (Kucera International Inc., Ohio, USA), yet we were not able to get the exact coverage we wanted for our irregularly-shaped community area since the cost of the maps prohibited us from buying more than one map set.

The two evaluative GIS maps that were created in this project reveal a high degree of consensus on likability and dislikability and suggest ways to improve the appearance of Pilsen. For instance, people repeatedly commented on the high density, the lack of parking, and the lack of trees and the preponderance of vacant lots. In particular, residents voiced strong dislike for spots on the north and east borders of the community, where railways, car yards, lumber yards, a garbage dump, and sheds are located. Given the high visibility of the places that do appear on the maps, such as locations along the Chicago River, changes in their appearance could have a strong effect on the public image of Pilsen.

Since the maps provide written evaluations for each point, they should be useful as the planning team proposes directions for improvements. The respondents also provided some helpful comments for improving the computer interface. They wanted the familiar landmarks to be marked to give them reference points when looking at the aerial map. Others commented that the aerial map was most helpful for highlighting large, general issues in Pilsen, such as land use and traffic flow; a different kind of map interface might be better for addressing design issues such as improving the visual appearance of the retail core. The planning team hopes to incorporate this input into the next stages of the Pilsen community planning process.

III. Ways to Improve and Expand the Tool

1) Map Design We are continuing to work on ways to develop the "optimum" map design/survey interface and are exploring how different map designs could lead to different kinds of information. We are asking questions such as, Are aerial photographs the best? What about color photos? Do we need more or less detail on the maps? Should we make the maps simpler? The design must facilitate and enable easy map reading. Planners here must realize that many people do not have a great deal of experience reading maps. It will be important to experiment to find out how people can best relate to the maps. We assumed in these projects (after some quick tests) that aerial photographs work well for people. But what about some cartoon maps with some annotation for major roads and businesses? Would these kinds of maps facilitate faster orientation to lay people or the novice? As we continue to develop this tool, we are consulting with Human-Computer Interface specialists on the UIC campus. They are advising us on how to better design the interfaces so that people can do the surveys intuitively with minimum misunderstanding and with minimum instructional materials. Further testing will be needed to refine the tools.

2) Greater Interactivity A future possibility that we have actually been testing is to have the survey tool return feedback to the respondent. This feature would allow the participant to provide input on the Web-based map about a question of public interest, and then, after submitting the information, the program would immediately return a cumulative map to the participant that integrates the individual's opinions with the rest of the people who had previously completed the survey. Another option we are working on is having participants draw their own boundaries for the locations they wish to comment on. The program may also accumulate the drawings done by respondents and be made to superimpose these accumulated drawings upon a return map as a color or line gradient.

3) Larger Geographic Scale While a smaller survey area provides more precision, there may also be value in expanding the tool to cover an even larger geographic area. We envision the possibility of applying the tools on the scale of an entire city, such as Chicago. The great benefit of the Web is that it provides the possibility for conducting two-way communication with a huge number of individuals. A city planning department or planning agency could use this tool in innumerable ways. For example, a Web-map survey tool could conceivably be used to answer questions as mundane as "Should we put the stop sign or new fire house here?" answered by a few hundred persons, to the more cataclysmic, such as "Should we put the toll way extension there?" answered by several thousand persons. Further work would need to be done in exploring how these large file sizes could be easily navigated.

4) Survey Research Standards At the present time, Web surveys typically suffer from under-representation and sample bias. As this tool and similar tools are further developed and refined, it will be important to adapt them to conform with established survey research standards, while maintaining the privacy of persons, establishing standards of informed consent, and at the same time maintaining the integrity of systems and data. This constitutes a major challenge to the "social sciences side" of this project.

This project avoided some of these issues because we had a relatively small, targeted audience who had interest in doing the survey. In our case, we had a specific community that we have been working with for several years on a variety of projects. Therefore, a relationship between the survey and potential respondents was pre-established. Also, we worked with community organizations that promoted this project and facilitated the "customers." Community organizations made special efforts to encourage residents, through personal, social and political connections and advertisements. The University's team assisted in the Web survey by having research assistants and interns help community organizations in conducting the survey. They welcomed community members and residents (who had no access to the Web) in our college's computer lab and in the community's offices, oriented them to the project, clarified questions, and helped them in conducting the survey.

IV. Conclusion

Architects and planners are increasingly recognizing that local knowledge is a key element in developing appropriate and effective solutions to community design and planning problems. In an environment where computer and Internet technology are becoming mainstream, it is important that planners develop ways to harness these technologies to work more effectively with the public. Computer visualization and Web-surveys offer a new way to support and facilitate democratic decision-making (Sarajokoski, 1998) and (Batty, 1998).

In this project, we used a map-survey to engage public knowledge and to discover a community's evaluative image. Access was facilitated through existing community groups and community networks in a low- to moderate- income community. The resulting composite maps show the location and likability/dislikability of visual features, and the degree of consensus for the evaluations. These maps provide valuable information about the public's perception of their own city form. The vision of the project was that the maps would provide planners, researchers and community leaders with a framework for developing physical design plans for the Pilsen community.

For us as researchers at the University of Illinois we feel that we achieved considerable success on the technical side of the project. The Web map-survey introduces an innovative way to involve the public in returning spatial data as well as consuming it. The project demonstrates the possibilities for navigating interactive maps on the Web, including moving between different scales. However, our experience in this project is consistent with Sarajokoski's analysis. At this point, the survey results are not yet effecting concrete changes in the community, primarily for organizational and political reasons. We had a difficult time setting up meetings with the community organizations to discuss the project after conducting the survey. City of Chicago officials and Aldermen were invited to meetings to discuss results but did not attend. We had to resort to a "same place, same time" meeting to discuss the collected information and make decisions based on the data. There were conflicts in scheduling meetings, making sense of the collected data, balancing varied interests, building consensus, making decisions and charting an action plan. While the Web in this project demonstrated great potential for learning about the community's needs and desires, moving these opinions into concrete actions and making decisions as a result of this information has proven to be the greater challenge. Despite these difficulties, the results of the Web survey will be continue to be discussed in the ongoing university-community urban design workshops and will be used to inform proposals that emerge from the planning process.

Through our experience in this project, we have identified several important issues that must be addressed as research in the area of Web-based public participation expands. Among these issues are increasing accessibility to computers and the Internet across all income groups, addressing social science research standards, and making sure the public's input will translate into tangible changes. If these issues are gradually addressed and resolved, Web surveys may begin to be used widely to disseminate and to collect demographic, socio-economic, public-preference, and other data of public interest, and to make these data readily available for interactive public participation. The final hope is that political leaders will be listening as more and more people use their computers to spell out their preferences for their communities.

References

Al-Kodmany, K. (2000). Using Web-Based Technologies and Geographic Information Systems in Community Planning. Journal of Urban Technology. 7(1): 1-30.

Appleyard, D. (1976). Planning a Pluralist City: Conflicting Realities in Ciudad Guayana. Cambridge: MIT Press.

Batty, M. (1998). Digital Planning. In P. K. Sikdar, S. L. Dhingra, & K. V. Krishna Rao (Eds.), Computers in Urban Planning and Urban Management Keynote Papers (pp. 13-30). New Delhi: Narosa.

Evans, A., Kingston, R., Carver, S. and Turton, I. (July 1999). Web-based GIS to enhance public democratic involvement Paper presented at GeoComputation '99, Mary Washington College, Virginia, USA.

Lynch, K. (1960). The Image of the City. Cambridge: MIT Press

Nasar, J. (1998). The Evaluative Image of the City. Thousands Oaks, CA: Sage.

Rapoport, A. (1970). Symbolism and Environmental Design. International Journal of Symbology, 1, 1-10.