InformatIon ScIence Reference Part 9 pot

While there are many childcare resources in the Jefferson County information guides, there is little overlap between the childcare listed in the Jeffer-son County guides and the childcar

Figure 6 Education resources: Cognitive mapping vs resource guides

Figure 7 Health care resources: Cognitive mapping vs resource guides

County resource guides failed to provide adequate

information Except in education and childcare,

the resource guides fell far short of the number of

resources identified by the mapping participants

While there are many childcare resources in the

Jefferson County information guides, there is little

overlap between the childcare listed in the

Jeffer-son County guides and the childcare identified by

the cognitive mapping The GIS maps effectively

demonstrate such knowledge gaps

Second, there are a significant number of

resources in Denver County (east of Jefferson

County) that providers and clients identify

Rea-sonable accessibility to Denver County, as well as

lack of availability of the resources in Jefferson

County, likely accounts for this trend Building

a community-based SOC will require Jefferson

County to find ways to offer some of these services

locally, a challenge that will require developing

community partnerships to overcome the financial

constraints which the County faces

Third, opposite of the previous trend, Jefferson

County resource guides provide mainly Denver

locations for some types of resources, even though

the same resources exist in numerous places in

Jefferson County Available resources closer to

Jefferson County residents are a fundamental

component of SOC and, in this trend, require only

disseminating the information effectively, which

is a low-cost method to improve community-based

service delivery

Finally, there is a large disparity in knowledge

between clients and providers With the exception

of 3 of the 24 categories, Education, Recreation,

and Commercial Resources, the providers and

clients did not overlap significantly in knowledge

about resources Providers know more about

traditional resources such as other agencies or

governmentally-supported social services, while

clients know about resources of a less traditional

nature, such as churches, motels, and parks where

teenagers gathered to socialize and engage in

recreational sports activities Although these

informal resources are not referral services that

providers typically pass along to clients, they are important community-based resources to share with clients In creating a community-based SOC, providers need to be aware of the alternative methods clients use to meet their needs In some instances, this new information will lead to the creation of government/community partnerships

to more effectively and efficiently deliver services

In other circumstances, the additional knowledge

of resources will provide clients with options and/

or fill gaps in needs that traditional government and community providers cannot meet

Lessons LeArned

Several problems directly and indirectly related

to the GIS component of the project became apparent and required adjustments to the pro-cedures or accommodations to the expected output These include research procedures that are incompatible with social service agencies’ capacity, issues of client confidentiality, repeat rates, incomplete and/or inaccurate databases for coding resource locations, coding protocols, and mapping accuracy

First, as has been found before, many county and local agencies lack leadership that under-stands the value of GIS in policy decision-making (Greene, 2000; Nedovic-Budic, 1996; Ventura, 1995; Worrall & Bond, 1997) Hence, many agen-cies lack the technical ability to employ GIS and, consequently, also lack the understanding to work effectively and efficiently with the researchers Furthermore, because social service agencies typically do not have a GIS analyst on staff, data and map files have limited usefulness beyond the initial analysis as presented in the final report Finally, human service agencies have organiza-tional procedures that create significant barriers

in implementing research projects, barriers that need to be addressed in the project planning stages (Ventura, 1995) Jefferson County Human Services suffered from all three impediments

and was exacerbated by the high turnover of the

staff In the first year, two-thirds of the project

staff left By the middle of the second year, only

one person out of nine key project staff remained

Those who left included the project manager and

the principal investigator, both of who had been

replaced twice Within 18 months, none of the

people who conceptualized and wrote the HHS

grant were involved in the project Institutional

memory was wiped clean and new staff was

unfamiliar and wary of many components laid

out in the grant proposal, including the

untra-ditional resource identification method Higher

administrative support for the innovative project

waned, and “business as usual” reasserted itself

as the dominant paradigm It became clear that

the resource database developed through the

map-ping process would not be updated on a regular

basis and, perhaps, not disseminated throughout

the organization if left to Jefferson County The

CIPP sought out a more stable organization to

house the resource data, Colorado 2-1-1, with the

permission of the first project manager

Second, human service agencies as well as

educational institutions cannot share

client/stu-dent data This presents a significant research

barrier when the project requires participation

of these populations Ideally, individuals within

the organizations would have both the access to

the data and sophistication to manipulate the data

in accordance with standard research protocols

This is unlikely to be the case in institutions which

are financially strapped and lack the vision or

political will to invest in trained personnel and

needed research tools To ameliorate these

condi-tions, project planning must include agreed-upon

protocols for effectively and efficiently handling

confidential data

Third, unique to this project was the

cre-ation of a “repeat rate” to set a standard for data

density The 80% repeat rate was selected for

efficiency of resources, based on an

extrapola-tion of the average number of points per map

and time needed to code and enter the data for

each map Unknown was how many participants/maps were needed to reach the 80% repeat rate

in each of the 24 categories Initially, the CIPP recommended target was 450 participants This number was revised downward by Jefferson County Human Services to a maximum of 250 participants From the 247 actual participants, the 80% repeat rate was reached in only two of the

24 resource categories The average repeat rate was 55% across all categories, indicating that more than 250 participants were needed to reach 80% Whether 450 participants were ultimately required is unknown More importantly, did the lower repeat rate significantly affect the quality

of the project? Certainly, fewer resources were identified at the 55% rate; but 1,480 resources not

in Jefferson County resource guides were fied; not an insignificant contribution to building

identi-a more comprehensive sociidenti-al services

Fourth, in the process of coding the maps and sorting the data to find repeated addresses

or groupings by type of provider, and so forth, it was discovered that precise alphanumeric coding was critical With the large number of data fields (attributes) assigned to each participant, there were inconsistencies in some of the categories The data cleaning was more extensive than anticipated Future projects should utilize numeric coding in attributes to the fullest extent possible and develop strict alphanumeric standards for addresses, or-ganizational names, and other alpha fields.Finally, to find resource addresses, MapQuest and the Denver metro area phone book were used MapQuest was the most efficient method but had the most errors, as discovered when the address was imported into ArcMap A cross-check with the phone books corrected most of these errors.Nine percent of the mapping points were unidentifiable due to a combination of missing information in MapQuest and the phone book, and poor location information on the hand drawn maps The latter accounted for a greater proportion of the unidentified points, especially resources such

as neighborhood parks and unnamed resources

such as “soup kitchen.” Rather than rely solely

on participants naming the nearest cross streets

to such resources, the closest known commercial

entity should be identified This redundancy will

reduce missing data due to participant error in

naming streets

f uture t rends

While this project was limited to identifying

resources, spatial patterns of resource locations,

and knowledge gaps, the collected data can be

mined further More specific uses can be created,

such as a searchable Web-based provider resource

database and the identification of physical and/or

service areas with inadequate resources in relation

to socio-economic deprivation areas The latter

allows providers to demonstrate specific needs,

important for several reasons, including the pursuit

of future programmatic funding These specific

uses are described in greater detail as follows:

the Web-based database can be converted

into a tool for social service providers to

identify available resources and the most

accessible locations for clients (Worrall &

Bond, 1997) The end user (a case-worker)

would be able to search for particular

re-sources based on any number of criteria or

a combination of criteria For example, one

might enter necessary criteria such as Rental

Assistance Housing Resource located within

three miles of a given location that also

caters to Spanish-speaking clientele After

these attributes or criteria are entered into

the appropriate locations on the Webpage,

a list of all the resources or providers that

fit the criteria could be retrieved, similar to

the business name search feature available

through a site such as MapQuest Finally,

digital maps could be generated with driving

directions for the case-worker to print out

for the client It is also possible to map the public transportation routes to services

used to conduct comprehensive, able, and defensible needs assessments A social service provider administrator or grant writer could search the data described above in conjunction with Census data and the County’s client locations to reveal areas

quantifi-of need or areas quantifi-of excess (Bond & Devine, 1991; Worrall & Bond, 1997).6 A strategic plan could be developed to determine where

a new office or access point for a particular resource should be located to serve the great-est number of clients This type of spatial analysis based on quantifiable numbers and distances can be used to justify a particular course of action either for internal/external accountability or to acquire funding for vari-ous projects aimed at community resource and social service distribution

Acknow Ledg Ments

The author would like to thank April Smith, Department of Psychology, Colorado State Uni-versity, and Mary Tye, Department of Psychol-ogy, Colorado State University, for running the workshops and coding the data; David Wallick, Colorado Institute of Public Policy, Colorado State University, for conducting the GIS analysis; and Juliana Hissrich for providing administrative support to the project

c onc Lus Ion

Cognitive mapping combined with GIS analysis

is a powerful method for identifying community resources by providing: (1) a comprehensive database of existing services; (2) a basis to build communication networks and cooperation among government and community providers; (3) the

ability to create an efficient system that avoids

duplication of efforts; (4) an understanding of

the geographical distribution of resources; (5) the

identification of resources lacking in the county

and specific communities; and (6) knowledge

differences among diverse participant groups

The addition of 1,480 resource locations within

the seven study areas (only a portion of Jefferson

County) nearly tripled the number of resources and

services listed in the Jefferson County guides

Ultimately, service delivery in SOC is about

building partnerships across the multiple services

and bringing in new, even sometimes

untradi-tional, community partners Family involvement

is the key in this collaborative arrangement

Similar to untraditional community partners and

resources, families as partners do not fit easily

within current social service delivery structures,

values, and beliefs Recognizing, valuing, and

partnering with resource providers identified by

clients and community members is one important

step toward shifting practices Cognitive

map-ping with GIS provides a tool for taking the first

critical steps

r eferences

Bond, D., & Devine, P (1991) The role of

geo-graphic information systems in survey analysis

The Statistician, 40, 209-215.

Daniels, K., & Johnson, G (2002) On trees and

triviality traps: Locating the debate on the

con-tribution of cognitive mapping to organizational

research Organization Studies, 23(1), 73-81.

Evans, G W (1980) Environmental cognition

Psychological Bulletin, 88(2), 259-287.

Fridgen, J D (1987) Use of cognitive maps to

determine perceived tourism regions Leisure

Sciences, 9(2), 101-117.

Fulton, W., Horan, T., & Serrano, K (1997)

Put-ting it all together: Using the ISTEA framework to

synthesize transportation and broader community goals Claremont Graduate University, University

Research Institute, Claremont, CA

Greene, R W (2000) GIS in public policy: ing geographical information for more effective government Redlands, CA: ESRI Press.

Us-Hardwick, D A., Wooldridge, S C., & Rinalducci,

E J (1983) Selection of landmarks as a correlate

of cognitive map organization Psychological Reports, 53(3), 807-813.

Heagerty, P J., & Lele, S R (1998) A composite

likelihood approach to binary spatial data Journal

of the American Statistical Association, 93(443),

1099-1111

Hjortso, C N., Christensen, S M., & Tarp, P (2005) Rapid stakeholder and conflict assess-ment for natural resource management using cognitive mapping: The case of Damdoi Forest

Enterprise, Vietnam Agriculture and Human Values, 22, 149-167.

Hobbs, B F., Ludsin, S A., Knight, R L., Ryan,

P A., Biberhofer, J., & Ciborowski, J J H (2002) Fuzzy cognitive mapping as a tool to define management objectives for complex ecosystems

Environ-Jordan, T., Raubal, M., Gartrell, B., & Egenhofer,

M J (1998, July) An affordance-based model of

place in GIS In Eighth International Symposium

on Spatial Data Handling ’98 Conference ceedings, Vancouver, BC, Canada (pp 98-109) Kathlene, L (1997) 29 th street greenway corridor citizen survey panel: Results of mapping exercise, phase 3 Minneapolis, MN: University of Min-

Pro-neapolis, Humphrey Institute of Public Affairs

Kathlene, L., & Horan, T (1998) GIS survey of

29 th street corridor, Minneapolis, MN

Minneapo-lis, MN: University of MinneapoMinneapo-lis, Humphrey

Institute of Public Affairs

Longley, P A., Goodchild, M F., Maguire, D J.,

& Rhind, D W (2001) Geographic information

systems and science New York: John Wiley and

Sons, LTD

Lynch, K (1960) The image of the city

Cam-bridge, MA: MIT Press

Magana, J R., & Norman, D K (1980)

Meth-odological inquiry into elicitation procedures:

Cognitive mapping and free listing Perceptual

and Motor Skills, 51(3), 931-934.

Horan, T., Serrano, K., & McMurran, G (2001)

GIS for livable communities: Examiniation of

community perceptions of assets, liabilities and

transportation improvements San Jose, CA: San

Jose University, Mineta Transportation Institute,

College of Business

Moeser, S D (1988) Cognitive mapping in a

complex building Environment and Behavior,

20(1), 21-49.

Nasar, J L (1988) The evaluative image of the

city Thousand Oaks, CA: Sage Publications.

Nedovic-Budic, Z., & Godschalk, D R (1996)

Human factors in adoption of geographical

infor-mation systems: A local government case study

Public Administration Review, 56, 554-567.

O’Laughlin, E M., & Brubaker, B S (1998)

Use of landmarks in cognitive mapping: Gender

differences in self report versus performance

Personality and Individual Differences, 24(5),

595-601

O’Neill, M J (1991) Evaluation of a conceptual

model of architectural legibility Environment

and Behavior, 23(3), 259-284.

Quaiser-Pohl, C., Lehmann, W., & Eid, M (2004) The relationship between spatial abilities and rep-resentations of large-scale space in children — a

structural equation modeling analysis ity and Individual Differences, 36(1), 95-107.

Personal-Reich, R M., & Davis, R (2003) Spatial tical analysis of natural resources (Tech Rep

statis-No NR512) Fort Collins, CO: Colorado State University

Sholl, M J (1987) Cognitive maps as orienting

schemata Journal of Experimental ogy: Learning, Memory, & Cognition, 13(4),

Psychol-615-628

Stroul, B (1996) Profiles of local systems of care

In B A Stroul and R M Friedman (Eds.), Systems

of care for children’s mental health (pp 149-176)

Baltimore: Paul H Brookes Publishing Co.Tolman, E C (1948) Cognitive maps in rats and

men Psychological Review, 55(4), 189-208.

Unger, D G., & Wandersman, A (1985) The importance of neighbors: The social, cogni-tive, and affective components of neighboring

American Journal of Community Psychology, 13(2), 139-169.

Uzzell, D., Pol, E., & Badenas, D (2002) Place identification, social cohesion, and environmental

sustainability Environment and Behavior, 34(1),

26-53

Ventura, S J (1995) The use of geographical

information systems in local government Public Administration Review, 55, 461-467.

Worrall, L., & Bond, D (1997) Geographical information systems, spatial analysis, and public

policy: The British experience International Statistical Review, 65, 365-379.

Yoshino, R (1991) A note on cognitive maps: An

optimal spatial knowledge representation nal of Mathematical Psychology, 35, 371-393.

1 The project was supported by grant

#90CA1715/01, CFDA #93.570 from the

Federal Department of Health and Human

Services through Jefferson County,

Colo-rado

2 The term cognitive mapping is used for a

va-riety of techniques, including “fuzzy

cogni-tive mapping,” a technique that builds mental

maps of perceptions from focus-group and

interviews (Hjortso, Christensen, & Tarp,

2005; Hobbs et al., 2002) In this project,

cognitive mapping means hand-drawn

maps of tangible community resources and

locations, a geographical data collection

technique new to GIS

3 Nine percent of the mapping points could

not be accurately located and were dropped

from the analysis Of the remaining 89%,

two possible location errors could occur in

transferring the cognitive map information

into a database for ArcMap First, multiple

coders could use different alphanumeric

codes, thereby making the same resource

appear as a different resource To correct

this error, the data was cleaned by

conduct-ing sorts on multiple columns in the excel

spreadsheet to reveal unknown duplicates

For example, a search on “Research Name”

might find the same resource with

inconsis-tent address codes If the address did not

match exactly (e.g., one was coded with

“St.” and another coded with “Street,” the

coding was corrected to be consistent

Simi-lar searches were done on other categories

such as street address, street name, and zip

code The data was cleaned accordingly The

second error was from incorrect addresses

in the MapQuest and/or Dex directory The

Dex directory is the official metropolitan

phone and address directory and should have

a high level of reliability; however, the actual

reliability rate is unknown To correct for possible errors, all identified social services not in the Jefferson County resource guides (e.g., soup kitchens, English as a Second Language courses, support groups, etc.) were called to verify the address It was assumed that the Jefferson County resource guides had accurate information

4 All identified resources were provided

to Colorado’s 2-1-1 system, which is the national abbreviated dialing code for free access to health and human services infor-mation and referral (I&R) 2-1-1 is an easy-to-remember and universally-recognizable number that makes a critical connection between individuals and families in need and the appropriate community-based organiza-tions and government agencies Housing the data with 2-1-1 allows statewide access to resources and bi-annual updating to keep the information current Colorado 2-1-1 system

is the depository for the resources collected

in this project Web searchable database of resources can be found at http://211colorado.org/

5 CIPP provided Jefferson County with the ethnic enclave areas based on the 2000 Census The Asian communities fell out-side the project boundaries set by Jefferson County (see Figure 1) and, unlike Russians, Latinos, and Native Americans, Jefferson County did not request mapping with the Asian community

6 For example, it might be found that 65% of all users of a certain type of resource (this data would be collected by cognitive map-ping alone) live “x” number of miles away (analysis performed by the GIS system) from

a particular needed or frequently-accessed resource (gathered through cognitive map-ping and other sources)

Append IX

Only forty percent of the participants provided

demographic information, which limits the

ability to determine the gender, age, and

ethnic-ity/race of the participants However, there is no

way to determine the representativeness of the

sample on these traditional demographics since

the population characteristics are unknown

Table 2 Demographics of participants (n=100)

Demographic

characteristic All participants(n=100) Providers(n=19) Clients(n=72) Community Residents (n=9)

Number and percent

a concern For the identification of resources, a cross-section of the types of people who use or provide services and the geographical distribution

of their knowledge was most important, of which both criteria were met

This work was previously published in Emerging Spatial Information Systems and Applications, edited by B Hilton, pp

326-350, copyright 2007 by IGI Publishing (an imprint of IGI Global).

Chapter XLV

Collaborative Mapping and GIS:

An Alternative Geographic Information Framework

Edward Mac Gillavry

Webmapper, The Netherlands

Abstr Act

The collection and dissemination of geographic information has long been the prerogative of national mapping agencies Nowadays, location-aware mobile devices could potentially turn everyone into a mapmaker Collaborative mapping is an initiative to collectively produce models of real-world loca- tions online that people can then access and use to virtually annotate locations in space This chapter describes the technical and social developments that underpin this revolution in mapmaking It presents

a framework for an alternative geographic information infrastructure that draws from collaborative mapping initiatives and builds on established Web technologies Storing geographic information in machine-readable formats and exchanging geographic information through Web services, collaborative mapping may enable the “napsterisation” of geographic information, thus providing complementary and alternative geographic information from the products created by national mapping agencies.

Introduct Ion

Since the Enlightenment, mapping and the

pro-duction of geographic information have been

institutionalised: the map is the power At home,

maps were used as an instrument for nation

building as nation states emerged: a

legitimi-sation device (McHaffie, 1995) People learned

about their country and administrations needed

a tool to govern the territory Away from home, maps were an instrument for colonisation, when Africa and Asia were split among the European nation-states

During the last few decades, there has been rapid democratisation of geographic information and maps Sawicki and Craig (1996) distinguish

three ways in which this movement is apparent

First, the locus of computing power and data access

is broadening Second, the level of skills to turn

raw geospatial data into geographic information

has become less demanding Third, the locus of

applications has moved closer to the citizenry

Geographic information systems moved from

mainframes and the UNIX operating system onto

personal computers and the Windows operating

system From research and government, GIS

spread into the business sector The PARC Xerox

Map Server and Virtual Tourist brought maps to

everyone’s PC in the late 1990s, followed by online

map Web sites such as MapQuest and Multimap

In 1997, Brandon Plewe noted that “the Internet

holds promise for exponential increases in the

efficiency and effectiveness of the ways in which

we obtain, use and share geographic information

in all its forms” (Plewe, 1997) In July 2002, 7.1

million European users visited one of the many

online map Web sites (Nielsen//NetRatings,

2002) Google Maps, introduced in February

2005, reached almost 1.7 million visitors in that

month (Buchwalter, 2005)

Although maps are more widely used than ever,

the production of geographic information, and

especially mapping, is still highly concentrated

among national mapping agencies and the GI

in-dustry But this oligarchy is soon to be dissolved,

for we see the third aspect of the democratisation

of geographic information–the locus of

applica-tions moving closer to the citizenry–becomes

apparent now that location-aware mobile devices

are coming within everyone’s reach GPS units

are not only available to surveyors anymore, as

cheaper devices are sold for outdoor recreation

Also, small GPS antennae can communicate

with other devices over Bluetooth, and there

are already mobile phones and personal digital

assistants (PDAs) for the consumer market that

have GPS-chips built in

At the same time, digital maps have become

portable Various mobile phone operators have

started to deliver location-based services to mobile

devices Mobile phones come with route planning applications, thus making in-car navigation sys-tems redundant Maps are not only delivered to the desktop, but also to mobile phones and PDAs, requiring new visualisations as the screen size, resolution, and use patterns differ significantly Collaborative mapping is an initiative to col-lectively create models of real-world locations online that anyone can access and use to virtually annotate locations in space (McClellan, 2003) The value of the annotations is determined by physi-cal and social proximity, the former expressed

in distance, the latter in “degrees of separation.” Thus, the informational value and the pertinence

of spatial annotations is not only dependent on physical distance, but also dependent on the trust relationship between individuals or groups

of people through social networks: the “Web of Trust” (Espinoza, Persson, Sandin, Nystrom, Cacciatore, & Bylund, 2001)

However, there is a discrepancy between cal and social proximity Privacy and personal freedom become highly important issues when one’s location is related to their social behaviour

physi-On the other hand, the fear of surveillance that accompanies positioning is already gradually reducing in society (Ahas & Mark, 2005) Fur-thermore, this discrepancy can be mediated by users themselves by storing annotations and tracks locally, thus creating distributed repositories, and by explicitly setting the level of privacy on each of these annotations and tracks Finally, users themselves remain in control of their so-cial identification–their preferences and social network–while they make use of collaborative mapping services, whereas, for example, the so-cial positioning method aggregates these social characteristics to study the space-time behaviour

of society (Ahas & Mark, 2005) Collaborative mapping services are therefore less pervasive in the privacy of their users because users negotiate the trade-off between the benefits of the service and their privacy concerns

Collaborative mapping is to location-based

services (LBS) what blogging is to the Web: a

mechanism for authoring one’s own

location-based experiences (Bleecker, 2005) The notion

of “social proximity” as a measure of relevancy

of geographic information sets collaborative

mapping apart from corporate LBS-providers

Requesting the nearest Italian restaurants, the

service not only takes into account the distance

to the restaurant from the current position of the

user, but it also navigates his/her social network to

find a suitable result People are willing to travel

farther because the particular restaurant is a

favou-rite haunt of one of their relations Furthermore,

geographic information is not merely broadcast

to users, but users can actively contribute to the

service: geographic information flows back to

the service A collaborative mapping service

then works like a spatially enabled notebook or

message board, depending on the privacy settings

people attach to their postings

Collaborative mapping may even become a

vehicle for community-based participatory

plan-ning As users share their thoughts, opinions, and

ideas about physical places, local knowledge,

community needs, and specific social histories

may thus be appreciated and incorporated into

the process (Harris, Weiner, Warner, & Levin,

1996) In this way, the necessary information

is produced in an operation whereby the local

knowledge arising from the social narratives, the

mental map, is converted into data within a GIS

for research and policy formulation

This chapter aims to describe the various

developments that have led to the necessity and

feasibility of collaborative mapping Based on

current practice, it goes on to describe an

ap-proach for an alternative geographic information

infrastructure that draws from collaborative

mapping initiatives and builds on established

Web technologies

bAckground

f ree g eographic data

Collaborative mapping as a practice to create geographic data was born out of the lack of af-fordable and copyright-free geographic data Governments spend billions to create large-scale geographic data to fulfill their civil and military responsibilities, and only governments can afford the hardware and software, as well as the labour costs of trained personnel involved in creating large-scale geographic data (Harris et al., 1996) Nevertheless, geographic data created by US government agencies such as the US Geological Survey (USGS) and the Census Bureau is avail-able copyright free and made available at the nominal cost of copying it, whereas in most other countries, geographic data has to be paid for and

is subject to strict copyright laws

Since the market position of national mapping agencies (NMAs) is changing rapidly, their bud-gets being cut and status changed, their geographic databases have become their main capital assets

In the new economy, the digital elevation model and digital line graphs are the new currency (McHaffie, 1996) To survive in their new role, they have to protect their assets with copyright restrictions and inhibiting pricing structures Whereas other government agencies and academic institutions in these countries obtain geographic data at a discounted rate, nonprofit organisations, (small) businesses, and individual citizens have found these protective measures inhibiting Not only pricing and copyright put up barriers As the

GI industry is evolving rapidly, patents have been granted to various key players in the market This practice is a delicate trade-off between encour-aging innovation and encouraging monopolies Collaborative mapping has therefore become

a practice to create copyright-free geographic data for free and to bring together existing freely available sources of geographic data

Positioning Technology

There are various developments that have

facili-tated the viability of collaborative mapping First

of all, the democratisation of mapping has not only

brought maps to everyone’s desktop or hand-held

device, the availability of positioning technology

contributes to this trend as well GPS devices

used to be the domain of surveyors and geodetic

engineers, but nowadays there are many cheap

GPS devices available for outdoor enthusiasts,

which become smaller and smaller As of 1 May

2000, selective availability has been turned off

so public users can locate themselves using GPS

with “military precision.”

With Bluetooth, separate GPS devices

commu-nicate with palmtops and mobile phones Recently,

there are palmtops and even mobile phones that

come with a built-in GPS antenna Furthermore,

the Enhanced 911 (E-911) US federal mandate

requires mobile phone operators provide precise

location information of mobile devices within 50

to 300 meters The ability to establish one’s

loca-tion is unprecedented The data that is obtained

from location-aware devices indicates the actual

location and movement of people, with increasing precision in real-time (Ahas & Mark, 2005).These location-aware devices turn everyone into a producer of geographic information, thus bridging the cartographic labour process Ever since the first differentiation of cartographic labourers that occurred with the invention and adoption of the printing press, when compila-tion and reproduction of the map were separated (McHaffie, 1996), mappers are “slogging” into the messy reality of the field again in order to produce the “map.”

Location Awareness

Publishing on the Web has changed over the last few years There is no need to have your own Web server or even know how to code HTML: nowadays everyone can publish on their Weblog These Web based publications consist primarily

of periodic articles in reverse chronological order Admittedly, most early Weblogs were updated manually, but tools to automate the maintenance

of such sites made them accessible to a much larger population

Figure 1 Various trends drive the uptake of collaborative mapping

Not only Weblog authors (bloggers) writing

about similar topics started to form communities,

but there also started to appear listings of Weblogs

that were from the same country (e.g., the now

defunct Gblogs) or even located in the same city

(BrightonBloggers) The next step came when

NYCBloggers (http://www.nycbloggers.com/)

and London Bloggers (http://londonbloggers

iamcal.com/) started indexing bloggers by their

nearest subway stop or tube stop respectively

Mapping the virtual space of the blogosphere on

to the physical one, these listings put a sense of

place back into cyberspace

In a final step, the positional accuracy increased

from country, town, and subway stop to the level

of geographic coordinates, while the geographic

extent increased to cover the world with GeoURL

(http://geourl.org/) Bloggers are encouraged to

add HTML meta tags to their Web pages to inform

visitors about their location:

<meta name=”ICBM”content=”52.09022,5.08159”>

<meta name=”geo.position”content=”52.09022;5.0

8159”>

<meta name=”DC.title” content=”Webmapper: what

the map can be.”>

GeoURL then spiders the Web and indexes

Weblogs that have these meta tags, and puts pins

on the world map indicating the locations of all

the Weblogs Some bloggers travel regularly, and

Weblogs can be updated using mobile devices,

so-called moblogging Location thus becomes

another means to structure Weblog entries Also,

bloggers frequently write about locations they

are visiting Developers have created extensions

for popular blogging tools such as Movable Type

for authors to attach a location to their Weblogs

and individual postings, and to keep track of

their travels, nearby Weblogs, and the locations

of their visitors Other tools such as IndyJunior

(http://www.bryanboyer.com/indyjunior/) and

World66 (http://www.world66.com/) allow

bloggers to show their travels and the countries

visited The World as a Blog (http://brainoff.com/geoblog/) shows the locations of recently updated Weblogs All in all, Weblogs have certainly raised the awareness of location as a means to structure information sources

Online Communities

Online communities have been around since the start of the Internet More and more, it is not only important to advocate that you belong to a certain community, but also who you know in various communities Search engines rate Web sites based upon the number of hyperlinks that link to a Web site Online social networks started appearing

in 2002, when the term was used to describe the means of networking in virtual communities, and became popular in 2003 with the advent of Web sites such as Friendster, Tribe.net and LinkedIn While Friendster allows members to chart their network of friends, LinkedIn aims at profes-sionals to keep track of their business contacts The popularity of these sites rapidly grew, and major companies have entered the Internet social networking space

Whereas these online social networks require visitors to subscribe to a particular network before using the service, and people end up subscribing

to all these Web sites, thus becoming a member

of all separate networks XFN and FOAF are distributed networks People define their own networks and can identify their online friends and the nature of their relationship As FOAF is based on RDF, it can be mixed with other RDF vocabularies, so resources can be shared among networks of people Thus, sharing information among a group of people based on the nature of relationships is a feasible thing

Open Source GIS

The final trend contributing to the birth of laborative mapping is the adoption of open source

col-in the field of GIS The IT col-industry has a long

history of open source projects, from

operat-ing systems, Web servers, and programmoperat-ing

languages to applications Only recently, the GI

industry is catching up, with open source

data-bases that support spatial features, for example,

MySQL and PostGIS There are several libraries,

for example, to convert between common spatial

data formats, or to change projections that are

included in various open source projects GRASS

has become a well-known open source desktop

GIS For online GIS applications, there are

vari-ous open source Web-mapping servers such as

Deegree, GeoServer, and the UMN MapServer

Web clients such as Chameleon and ka-Map

al-low developers to build compelling interfaces

to work with maps in your Web browser Most

of the architectures are built around OGC (open

geospatial consortium) standards such as the Web

Feature Server and Geography Markup Language

specifications

Online vector-based mapping used to be the

domain of proprietary plug-ins from leading

GIS vendors As the XML-based scalable vector

graphics (SVG) file format gains popularity,

vec-tor-based mapping is not anymore the monopoly

of the GIS industry and government agencies

Many individual Web developers have started

to create compelling vector-based Web-mapping applications

With geographic data being transferred based upon open standards and clear interfaces, it is easy for developers to combine various data sources and make interesting Web-mapping applications, with no resources spent on buying software

c o LLAbor At Ive MAppIng

Several projects are underway to collect graphic data for the creation of base maps Amsterdam RealTime (http://www.waag.nl/re-altime/), conceived as an art project, has been a trigger for many people to realise the potential of collaborative mapping As part of an exhibition

geo-on historical maps of Amsterdam in late 2002, visitors were invited to participate and take a GPS device with them for a week During that time, their locations were uploaded in real time to a Web server and beamed onto a screen in one of the exhibition rooms Visitors of the exhibition saw a new map of Amsterdam gradually appear

as the individual GPS tracks joined up and started

to overlap

Figure 2 The iterative collaborative mapping process consists of four steps

These base maps typically consist of road

networks in urban areas that have been traced

moving around the area with a GPS device Most

community mapping projects seem to follow

four steps: collection, editing, augmentation,

and distribution (Walsh, 2005) In the following

paragraphs, these steps of community mapping

projects are discussed Although the paragraphs

are provided in this sequence, community

map-ping is an iterative process The steps are regularly

iterated as new data becomes available, and these

steps may take place concurrently

c ollection of base Map data and

gps t racks

There are several ways to kick-start a

collabora-tive mapping project One approach is to scan

and georeference out-of-copyright paper maps or

global satellite imagery, and incorporate these as a

background to provide a geographic context for the

project The advantage of global satellite imagery

over scanned paper maps is the flexibility in the

projection Opting for scanned paper maps, there is

not this flexibility, as the map image may become

distorted when changing the projection

Another approach is to use readily available

vector-based geographic data, having the

flex-ibility of projecting the data while

maintain-ing the sharpness of vectors For example, the

TIGER/Line data from the US Bureau of the

Census provides a good source for geographic

data to start a community mapping project As

discussed previously, copyright restrictions may

prevent projects from incorporating these sources

For the Mumbai Free Map project (http://crit.org

in/projects/gis), the Collective Initiative Research

Trust (CIRT) builds the base map from the

cu-mulative basis of the urban research, design, and

development studies and interventions conducted

previously by CIRT Thus, CIRT develops an

open-access spatial data infrastructure, and a

set of simple tools and applications localised in

Indian languages, for knowledge transfer and

participatory urban planning by communities and citizens in Mumbai

The third approach to collect base map data is the most labour intensive This low-level approach requires a large group of community members to travel the area to be mapped with a GPS receiver, and follow the tracks, paths, and streets to trace the network On their travels, members may add additional waypoints for landmarks in the area Returning from these surveys, the community members upload their GPS tracks over the Web

to a central server, or equally make their tracks available online for others to harvest Previously collected GPS data created while engaging in a game of geocaching or as a form of GPS-drawing, can be equally made available for integration

In the United Kingdom, GPSdrawing (http://www.gpsdrawing.com/) uses a GPS receiver to make pictorial drawings while moving around Conceived as an art project, base mapping is not the objective here, but the accumulation of GPS tracks makes a nice base map Geowiki (http://www.geowiki.com/) is a cartographic variation

of the interactive encyclopaedia Web site, pedia GPS tracks are converted into PostScript files that are merged, using a drawing package, into a GIF file that is put on the Web site Apart from showing the road networks in some of the larger towns in Britain, the Web site allows visi-tors to enter information about pubs and shops, with personal reviews

Wiki-Whereas Geowiki uses a rather onerous production process to convert the GPS tracks into a base map, recent initiatives build on the MapServer Open Source WebGIS package to serve online maps on the Web The MapServer was set up in compliance with standards from the OGC, particularly the Web feature server with transactional extension (WFS-T) to support not only serving maps on the Web, but also to edit the maps remotely The London Free Map (http://uo.space.frot.org/freemap/) initiative finds pedestrians, bikers, and skaters in London volunteering to collect GPS tracks Of course,

these data are difficult to integrate with the road

networks for car navigation, as these tracks are

not representative for cars However, as personal

navigation becomes more important, it is useful to

have navigation systems that include cycle tracks

and walk paths The London Free Map cooperates

closely with the Open Street Map (http://www

openstreetmap.org) initiative Steve Coast bikes

around London carrying a GPS device (Dodson,

2005) Coming home, he downloads the GPS

tracks onto his laptop to subsequently make them

available online to share the tracks with others

Street by street, a copyright-free map of London

is built

Editing Waypoint Data

This step involves the actual creation of the base

map from the raw GPS tracks previously

col-lected First, it has to be decided which tracks

or individual GPS points to consider for

inclu-sion in this step Next, geometric and statistical

algorithms have to be applied to decide when

several very close lines are actually the same

line and thus need to be averaged together, or are

actual parallel segments Intersections have to

be modelled into junctions at which tracks have

to be split in segments to ensure topology Other

points may belong to traffic situations such as

squares, roundabouts, and parks

Quantitative mechanisms relying on geometric

algorithms, or qualitative mechanisms requiring

human intervention, may decide which GPS tracks

are relevant at this stage in the process Geometric

algorithms may filter out GPS points that have

a significantly different accuracy from the rest

This may be important in areas with bad satellite

reception such as urban canyons, or at the start of

a GPS track due to cold start: sometimes it takes

rather long for a GPS device to establish

connec-tion with an appropriate number of satellites

Equally, GPS points, tracks, or repositories can

be included or excluded from the GPS tracks that

are the input for the base map based on

qualita-tive mechanisms (Shorter, 2005) There may be

a hierarchy of users, with trusted users given a higher moderation status based on the quality of their contribution to the community, to review contributions To ensure community members remain interested and help reviewing, RSS feeds

of latest updates and additional repositories of GPS data can be made available to trigger people to help moderate the GPS tracks Moderators and com-munity members can discuss among themselves about the value of GPS repositories on discussion forums that are provided at a central server, or that can be left as comments on distributed repositories and be amalgamated using RSS readers

As a save approach, GPS data is not able for postproduction editing unless one com-munity member has reviewed the contribution The strength of collaborative mapping is that relationships within the community can be made explicit Therefore, members may only want to include GPS tracks and landmarks in their base map that were left by people within their personal network of friends Finally, previous moderations

avail-of the GPS data remain available for revision avail-of the individual member, and they can roll back moderations on request, just like a wiki

Edelkamp and Schrödl describe a “map ment process” to create a traversable road network from GPS tracks (Edelkamp & Schrödl, 2003)

refine-In the context of normalising and synthesising GPS tracks, the road segment clustering stage and the road centreline generation stage of the map refinement process are relevant The road segment clustering stage identifies segments that are common between different GPS tracks and locates intersections of these segments, based on the contiguity information and temporal order of the GPS points The road centreline generation cre-ates a geometric construct that captures the road geometry The centreline is generated from a set

of GPS points using a weighted least-squares fit The relative weight of each GPS point is derived from the measurement error, which is usually available from the GPS receiver

Annotation of Geometry with

r eal-w orld semantics

This step ensures the geometry of the network

makes sense for people that use the network to

locate themselves and navigate around the area

The driving restrictions at one-way streets and

turning restrictions at the intersection of segments

have to be added by ground truthing: surveying

the actual location to interpret the data correctly

Surveyors go out with a printed base map to

an-notate and upload annotations at a later time, or

may well carry PDAs connected to a WFS-T map

server and add annotations on the fly

The Downhill Map of Bristol

(http://locate.ira-tional.org/bristol_map/) is a map for

skateboard-ers, cyclists, and walkers Artist Heath Bunting

and the Bristol Skate Survey group surveyed the

streets of Bristol, the road surface quality, and

the locations of fruit trees Like Geowiki, the

Downhill Maps of Bristol incorporates landmarks

along the street network created This initiative not

only simply records the longitude and latitude, but

also maps the attributes of the map object such as

height and road surface quality From this data, a

functional online route planning service can be

created for cyclists, skateboarders, and skaters,

which no national mapping agency would be able

to offer (Albert, 2004) The Dutch GeoSkating

project (http://www.geoskating.com/) aims to

cre-ate a similar map of the outskirts of Amsterdam,

uploading the GPS tracks and notes on road surface

quality to a Web server in real time

Not only navigation attributes such as road

type, turn restrictions, and one-way streets are

col-lected in this step, but also toponymy is important

here: the geographic names of roads, streets, and

landmarks In the context of collaborative mapping

projects, geographic names are collected using

the notion of “consensus-based location,” a term

coined by Alasdair Turner at the PLAN meeting

in London, early 2005 The name of a location is

established through consensus among a group of

people It is not determined a priori, but evolves

as people in a community add new annotations

in response to the needs of the moment

For example, in the GeoNotes service noza et al., 2001), a “location tag” determines the exact location of an annotation in the service in case the accuracy of the positioning technology of either the creator or the reader is too coarse Thus, other users can still verify the correct location of the note without having to rely on the accuracy

(Espi-of the positioning technology Furthermore, the positional accuracy of the location tag may evolve with improving positioning technology, and de-pending on what labels are defined

The communal character is emphasised as these location tags are shared in the community The set of location tags is stored in the service for other people’s perusal Upon tagging a note with a location, other notes in the vicinity are scanned for their location tags, and presented

as a list of options for location tags, sorted cording to popularity The creator selects a tag from the list or creates a new location tag This service may be bootstrapped with predefined labels taken from gazetteers, the yellow pages,

ac-or tourist databases

This strategy not only applies to point tions, but to streets segments and areas as well The Neighborhood Project (http://hood.theory.org/) uses the TIGER/Line US Census data for mapping city neighbourhoods based on the col-lective opinions of Internet users Addresses and neighbourhood data collected from housing posts

loca-on Craigslist are translated into latitude and lloca-on-gitude values and drawn on the map The blobs

lon-on the map thus give a notilon-on of the spatial extent

of neighbourhoods derived from the collective mental map of Internet users

Distribution

Once the geometry of the street network and the locations of landmarks have been created, they have to be made available to a wide audience, without copyright or financial restrictions In the

previous steps, various projects were highlighted,

with the aim to build geographic databases along

While they may coexist to make up a distributed

repository of GPS tracks, it may be advisable

to cooperate, to ensure GPS tracks from these

repositories can be easily incorporated into any

of the other repositories,and base maps can be

exchanged without any problem Therefore, it

is advisable for these projects to work together,

but also to transcend boundaries and work with

other initiatives that may benefit from geographic

data, for example, civic initiatives Already, the

London Free Map, Geowiki, and OpenStreetMap

are cooperating to build a street network for

London together

There are various sets of geographic data

cre-ated as part of a collaborative mapping project The

scanned and georeferenced base maps, the global

satellite imagery, incorporated as background,

can be made available, as well as the geographic

vector data For example, the London Free Map

initiative hosts various BitTorrents to distribute

free geographic datasets in a cost effective way

Also, the individual GPS tracks collected during

the project can be made available as GPX files, a

lightweight XML grammar to express GPS data,

or encoded using the geography markup language

(GML) published online as a Web feature service

(WFS) Also, for applications that do not require

the original vectors, raster images of the street

network can be published as a Web coverage

service (WCS) that can be used as tiles to plot

other data on top of

Challenges and Opportunities

Sawicki and Craig (1996) identify two

contra-dictory currents in information technology that

equally apply to collaborative mapping services

On the one hand, there is a movement towards

more accessibility by more users: online mapping

has experienced a steady uptake, and more and

more people turn to Web maps to find their way in

the real world Also, mobile phones and PDAs are

becoming new platforms for delivering geographic information Since the mid-1990s, more than 50%

of the citizens of the developed countries have become owners of mobile phones, while mobile communication has also been remarkably success-ful in poorer countries (Ahas & Mark, 2005) On the other hand, there is a movement towards less accessibility Although more users have access

to more data and information, the entry cost of the education, training, capital cost, time, and experience of the users to take advantage of in-formation technology is becoming much greater (Sawicki & Craig, 1996) As an example of this movement, Ahas and Mark (2005) write that the uptake of LBS has slowed down, as the services proved too difficult for the public to obtain, and users have difficulties imagining the potential of spatial information For collaborative mapping initiatives to become successful, it is advisable

to start off with the lowest common technical denominator, but to design and develop architec-tures to cater for more advanced technology in the future as location-aware devices become cheaper Already, projects such as the London Free Map use Web-based open source GIS technology and cheap GPS-devices rather than location-aware mobile technology

Since the information value and pertinence

of collaborative mapping services is ised not only by physical proximity, but also by social proximity, participation becomes another challenge when only a small group has access

character-to location-aware, mobile devices Especially in the case of applying collaborative mapping as a form of community-based participatory planning (Harris et al., 1996) or as a means to obtain data for the social positioning method as input for urban planning (Ahas & Mark, 2005), one has to

be aware that it is only based on the space-time behaviour and mental map of only a small group within society, with specific social identifica-tions The dependency of collaborative mapping services on social proximity may well expect a wider participation, as indicated by the popularity

of online social networking tools, thus possibly

pushing the uptake of LBS in general

Current collaborative mapping initiatives

rely on GPS-devices to capture road network

geometry and open source GIS technology to

serve maps and edit the raw data The ubiquity

of location-aware devices will increase over the

next few years Not only will the pool of possible

volunteers to participate in collaborative mapping

projects therefore expand accordingly, but the

collaborative mapping process will also benefit

from these technological advances as more raw

positional data is being captured with increasing

accuracy and in real time (Ahas & Mark, 2005)

Finally, the distributed nature will be reinforced,

as more collaborative mapping services will

de-rive spatial annotations from more repositories,

and will have to take into account wider social

networks when establishing the pertinence of

these annotations

Future Trends

Ben Russell (2003) suggests a “street server” as

a vehicle for distributed collaborative mapping

Street servers can be thought of as a computer

network with both private and public functions

for a community of residents of a street In the

context of collaborative mapping, these street

servers operate as a GPS tracks database that

can be linked with other street servers to form a

map covering a wider area, as far as GPS tracks

match up Thus, mobile devices ask stationary

street servers about their position and how to

travel around the neighborhood This network

approach contrasts with the current practice of

serving GPS tracks from a single, centralized

Web server

As geographic information is created

col-laboratively, people upload their annotations to a

central repository, or store their annotations and

tracks on their personal Web sites, thus creating

a distributed framework for sharing geographic

information With increasing storage capacity of

mobile devices, people may even put their GPS tracks on their location-aware mobile devices and share the data with their peers in their social net-work, just as digital music files are shared online

or made available as podcasts Ed Parsons, CTO

of the Ordnance Survey, spoke in an interview about the “napsterisation” of geographic informa-tion, claiming that “the experience of the music industry may also apply to geographic informa-tion” (Westell, 2003)

Thus, as people roam around the city, their mobile devices pick up blocks of geographic data

to locate themselves and navigate the area rounding them That geographic data represents the street network and landmarks of the city blocks where they find themselves, as derived from col-laboratively collected GPS tracks

sur-c onsur-c Lus Ion

Collaborative mapping is the latest development

in the democratisation of geographic information The proliferation of positioning technology, loca-tion-awareness, online communities, and open source GIS technology have paved the way for people being able to contribute geographic data and access localised information Various collab-orative mapping projects are currently collecting GPS tracks to derive street network data The framework of steps described in this chapter was based on experiences from these projects, and will stimulate the exchange of geographic information among these projects Hopefully, collaborative mapping will become a complimentary input for corporate location-based services

r eferences

Ahas, R., & Mark, Ü (2005) Location-based services – New challenges for planning and public

administration? Futures, 37(6), 547-561.

Albert, S (2004) Crossing the line between

mapping and mapmaking: Artistic approaches to

cartography Society of Cartographers Bulletin,

38(1-2), 41-44.

Bleecker, J (2005) A design approach for the

geospatial Web Retrieved February 2006,

from

http://www.oreillynet.com/pub/a/net-work/2005/06/07/geospatialweb.html

Buchwalter, C (2005) Integrated interactive

marketing:Quantifying the evolution of online

engagement Retrieved from

http://www.nielsen-netratings.com/pr/wp_050401.pdf

Dodson, S (2005, April 7) Get mapping The

Guardian Retrieved from http://www.guardian.

co.uk/online/story/0,3605,1453293,00.html

Edelkamp, S., & Schrödl, S (2003) Route

plan-ning and map inference with global positioplan-ning

traces In R Klein, H W Six, & L Wegner

(Eds.), Lecure notes in computer science 2598

Computer science in perspective: Essays

dedi-cated to Thomas Ottmann (pp 128-151) Berlin:

Springer

Espinoza, F., Persson, P., Sandin, A., Nyström, H.,

Cacciatore E., & Bylund, M (2001, September

30-October 2) GeoNotes: Social and navigational

aspects of location-based information systems In

Abowd, Brumitt, & Shafer (Eds.), Ubicomp 2001:

Ubiquitous computing, International Conference,

Atlanta, Georgia (pp 2-17) Berlin: Springer

Harris, T M., Weiner, D., Warner, T A., & Levin,

R (1996) Pursuing social goals through

particat-ory geographic information systems Redressing

South Africa’s historical political ecology In J

Pickles (Ed.), Ground truth The social

implica-tions of geographic information systems (pp

196-222) New York: The Guildford Press

Mac Gillavry, E (2004) Collaborative mapping:

By the people, for the people Society of raphers Bulletin, 37(2), 43-45.

Cartog-McClellan, J (2003, March 27) Get caught

mapping The Guardian Retrieved from

ry/0,3605,922337,00.html

http://technology.guardian.co.uk.online/sto-McHaffie, P H (1995) Manufacturing metaphors: Public cartography, the market, and democracy

In J Pickles (Ed.), Ground truth The social plications of geographic information systems (pp

im-113-129) New York: The Guildford Press

Nielsen//NetRatings (2002) Seven million ans look to the Web for maps, guides & directions

Europe-Retrieved from http://www.nielsen-netratings.com/pr/pr_020920_uk.pdf

Plewe, B (1997) GIS online: Information trieval, mapping, and the Internet Santa Fe:

re-OnWord Press

Russell, B (2003) Street servers – An outline

of an open source networked neigbourhood

Retrieved from servers/street-servers.htm

http://www.headmap.com/street-Sawicki, D S., & Craig, W J (1996) The ratisation of data: Bridging the gap for commu-

democ-nity groups Journal of the American Planning Association, 62(4), 512-523.

Shorter, C (2005) Bicycle map builder white paper Retrieved from http://bikemap.openearth.

com.au/docs/whitepaper/

Walsh, J (2005) London Free Map Retrieved from

http://uo.space.frot.org/?node=LondonFreeMapWestell, S (2003, July/August) The napsterisation

of geographic information GI News.

This work was previously published in Collaborative Geographic Information Systems, edited by S Balram and S Dragicevic,

pp 103-120, copyright 2006 by IGI Publishing (an imprint of IGI Global).

Introduct Ion

Recent studies have revealed that over 80% of

local governments in the U.S have locational

references in their data (Nedovic-Budic & Pinto,

1999) and a majority of local governments use

geographic information systems (GIS) technology

to manage spatial data, a trend often described as

a “growth surge”(Warnecke, Beattie, Cheryl, & Lyday, 1998) With the growth of Internet, there

is an increasing demand for location specific data and analytical solutions requiring GIS to locate

and integrate multiple databases This, in turn,

requires federal, state, and local government

agencies to develop capabilities so that their data

can interoperate For example, a real estate

entre-preneur, looking for a suitable location for a new

business, would require data that combines GIS

data with that of the local government’s zoning

and tax incentive areas Home owners and home

buyers, looking for information about

environ-mental hazards, can use e-maps that combine

data from several sources including the EPA’s

environmental data, and the HUD’s (Department

of Housing and Urban Development) housing

community programs (GAO, 2003) Similarly, a

water/sewer storm water utility company

evaluat-ing the feasibility of a new project to expand the

existing infrastructures in mountain areas may

need information about geomorphologic

forma-tions and associated potential landslide risk from

the local and federal government databases

Agencies in various levels of government rarely

coordinate the development of their applications

Consequently there is often redundancies and

duplications of data even within organizations

belonging to the same jurisdiction (Nyerges,

1989) Users often have to deal with proprietary

systems that require the understanding of the

systems’ native command language, input data

format, and output presentations The problem

is further compounded when there is a need for

communicating with more than one modeling

paradigms or when spatial analysis and

model-ing techniques are used in application areas for

which they were not necessarily designed In most

cases, users’ access, inference, and analytical

ability of spatial dataset and services are limited

by proprietary standards, platform dependence,

and incompatibility

In an e-government environment, simple

trans-actions can require intertrans-actions among multiple

resources possibly from different entities within

the government, and meaningful understanding

of system architectures and the service

compo-sitions Interagency transactions become simple

if the agencies involved in a transaction have homogeneous representation structures as well as the same discourse domain (Malucelli, Palzer, & Oliveira, 2006) A geospatial application can use business services with relative ease if it can un-derstand another application’s service descriptions and representations of workflows and information flows within and across organizations However, these representations become complicated when one needs to embed complex data structures and models into an application For instance, suppose

we are interested in a mobile commerce tion that would provide geospatial information as

applica-a prelude to completing applica-a business trapplica-ansapplica-action The transaction protocol for such an application would require access to and representation of geographic data and models These models themselves may require chaining of multiple services that depend

on service level description of geo-processing models, spatial geometries, spatial analysis, and implementation logic Typical query such as “Find the nearest Italian restaurant along the highway” could possibly be answered by chaining multiple services such as geocoding points of interest, integrating transport networks, creating dynamic segmentation of network, providing routing net-work, rendering cartographic information, and possibly converting text to voice It is possible

to envision integration and chaining of services

to provide higher levels of functionality if such services are distributed all over the enterprise and are accessible in a uniform standard manner (Peng & Tsou, 2003)

Methodological artifacts, techniques for rect description, and interpretation of resources, collectively known as the semantic layer (Vetere

cor-& Lenzerini, 2005), are pre-requisites to high level interoperability in a service-oriented envi-ronment High level or semantic interoperability

is of vital importance if collaborative business processes are involved (Padmanabhuni, 2004)

A complex collaborative process is often needed

to compose complex services like dynamic sualization and query processing of geo-spatial

vi-data in real time The representation of semantic

content of space and spatial knowledge has a

special significance in information exchange and

integration As representation of spatial features

are scale dependent, multiple representations of

spatial objects require contextual information of

the spatial reference domain that allows spatial

features to be integrated into application models

Therefore, typical business transactions involving

spatial data interoperability require processes

beyond the data translation or the conversion of

geometric primitives In collaborative business

processes involving multi-agent interactions, the

semantics of the complex spatial application

mod-els needs to be shared to infer or draw conclusions

from such annotations However, understanding

the collaborative aspects of multi-agent

interac-tion to solve complex spatial problems is still in

its infancy (Sikder & Gangapadhayay, 2002)

This article proposes a framework for a

se-mantic-level communication between geo-spatial

services in business processes and application

models The article presents an overview of

interoperability efforts with specific reference

to geo-spatial databases and application models

and reviews the feasibility of an ontology-based

spatial resource integration to combine the core

spatial reasoning with domain-specific application

models Existing industry standards and practices

in geo-spatial interoperability are identified

This is followed by a discussion of the role of

ontology in explicating the implicit semantics

of spatial data models and the need for

formal-ism in descriptions of spatial categories Use of

markup languages for spatial resource description

and the tagging of spatial ontology are

illus-trated Finally, a multi-agent based architecture

(OSIRIS-Ontology-Based Spatial Information

and Resource Integration Services) for semantic

interoperability of spatial data sets and models

is proposed The architecture is illustrated using

an application model that uses domain ontology

of urban environmental hydrology

g Is ser vIces In the

f r AMework of e-g overn Ment

As stated earlier, government agencies within the same jurisdiction rarely coordinate their ap-plication development as well as standardization

in data representation Making data available electronically then becomes a problem whether the potential user of the data is a government agency or otherwise A major initiative in the USA seeks to address this problem The National Spatial Data Infrastructure (NSDI) seeks to build

an organizational and virtual network to promote the sharing of spatial data among federal, state, regional, and local government agencies and the private sector The Federal Geographic Data Committee (FGDC) (FGDC, 2006) is tasked to develop a spatial data and metadata standard as well as to create data clearinghouse FGDC is also responsible for coordinating the development

of a “framework” data (Wayne, 2005) Similar initiatives have been taken by European commu-nities for developing a spatial data infrastructure (SDI) for e-government (Craglla & Signoretta, 2000) As a part these initiatives to share data and services among public and private agencies

as well as to achieve interoperability, various legal frameworks and standards have been insti-tuted The E-government Act of 20021 requires federal agencies to coordinate the development

of standard protocols for sharing geographic information to reduce redundant data collection and promote collaboration and the use of stan-dards (GAO, 2003) As a result, the role of NSDI and SDI is evolving from a simple data access facilitator to integrated service provider where complex geospatial services and functionalities can be chained in a distributed environment to meet user needs The framework of SDI and the role of e-government are viewed as a portal that can be used to access e-government services The geospatial one-stop portal, implementing the NSDI framework, seeks to coordinate and align geospatial data collection and maintenance across

all levels of government This portal envisions a

virtual repository of spatial data and Web services

and can provide support to local, state, and federal

programs as well as decision making at various

level of e-government operation The portal can

be used by users to access the following services

(Cummens, 2003):

1 Searching the geospatial one-stop network

for image and feature Web services;

2 Registering for notification when new or

updated data, maps, activities, and

refer-ences are added to system repositories;

3 Viewing metadata to determine if a product

is suitable for the intended use; and

4 Publishing (registering) map services,

im-ages, geographic data sets, geoservices,

spatial solutions, geographic, and land

reference materials

The geospatial one-stop portal can to be

ap-plied in the real world environment For example,

in a real estate business, the portal can perform

the task of a ‘broker’ to discover and chain land

information or cadastre services from various

nodes according to the specifications set by

ser-vice requesters (Radwan, Bishir, Emara, Saleh,

& Sabrah, 2005)

The application of Web services in

geospa-tial service management offers an opportunity

towards the local autonomy of databases Web

services can be used to dynamically query various

GIS layers while maintaining local agency level

independence in a distributed environment From

an organizational point of view, this may be very

appealing Local governments, such as counties

can independently collect and manage data and

still integrate information and services using

Web services A client of the local government

entity, for example, a transportation company,

can directly access the government’s base map

without maintaining its own dataset With

ap-propriate permissions, a client can also update

the government’s dataset from the client’s own

record An extended collaboration and partnership between multiple agencies and clients using Web services can provide opportunity to interoperate through open interfaces and communication protocols Typical geo-processing services may include data management tools like projection and transformation, topology manipulation, indexing, and spatial join

The interoperability issue in the context of e-government services is not limited to technical issues such as linking computer networks There

is a fundamental requirement to share and re-use knowledge networks and reorganize adminis-trative processes to better support the services themselves The key areas of interoperability

in need of consideration when implementing government services include the organizational, semantic, and technical issues (CEC, 2003, 2006)

e-At the semantic level, there is an imperative that the meaning of the information exchanged is not lost in the process of acquiring e-government services from all levels of government Thus, semantic interoperability entails seamless integra-tion of information and services from agencies at different levels whether they are local, regional,

or national This ensures organizational or local autonomy as well as global integration

However, there is growing skepticism whether the vision of geospatial one-stop’s objectives will

be fulfilled (GAO, 2003) and interoperability can be achieved The main obstacles are lack of standards and the requirement of a huge metadata While existing FGDC metadata provides content standards, they lack a framework for semantic annotation of the geospatial content and services The geospatial one-stop portal requires a semantic layer that would contain domain level specifica-tion of knowledge, mediator, and mapping agents

to coordinate among mapping components The semantic layer should be able to decompose users’ declarative requests, compose complex spatial services within the context of user-defined constraints

epIste Mo Logy of

Interoper AbLe Agents

In bus Iness processes

Epistemologically, the intrinsic state of an agent’s

reflection of a business process is partially

char-acterized by the internal schema (e.g., relational

or object relational) of the agent In addition, the

agent also offers manipulation methods that can be

used to interface with other agents through specific

access protocol Such autonomous agents can be

considered as a proxy who works collectively on

behalf of some higher entity having rich schema

or domain knowledge For example, in a B2B

e-commerce system, a cooperative mechanism agent

may negotiate with multiple suppliers, monitor

auctions, and infer tactical business strategies in

a competitive environment Agents may not only

involve in collaborative interactions to achieve a

common goal, they may also identify best partner

through negotiations (He, Jennings, & Leung,

2003) Business applications can then exploit

these agent benefits Specifically, a large complex

problem can be decomposed into subtasks that can

then be distributed among the agent

communi-ties specializing in solving domain problems

In a multi-agent environment, the internal state

of agent’s reflection needs to be translated to an

export schema to map conceptual correspondence

Therefore, multi-agent communication paradigms

require semantic level agreement Such semantic

interoperability eventually boils down to the

prob-lem of the identification of semantically similar

objects that belong to different components and

the resolution of their semantic differences (Yeap

& Handley, 1990) In this sense, the semantic

interoperability problem is somewhat similar to

the problem of schema integration in traditional

multi-databases

Different levels of business process

conceptu-alization have given rise to many domain specific

heterogeneous systems, which are often very

difficult to interoperate in applications In recent

years, there has been a significant level of

inter-est to study the role of an agent in system level interoperability by offering interface with special-ized systems that require agents to understand the native command language, input data format, and output presentations In such heterogeneous environments, agents require mapping one-to-one objects on a syntactic basis or measuring the semantic equivalence of objects However, when the representation is complex, for example, spatial databases, simple semantic translation is inadequate as it can result in semantic conflicts Three different types of semantic conflicts have been identified in spatial applications (Goh, 1997): confounding conflicts, scaling conflicts, and naming conflicts Confounding conflicts occur when information items have the same meaning but have different realizations Information items representing the same value but with different reference systems lead to scaling conflicts; and naming conflicts are generated when same items have different rules in different systems A typical example of spatial semantic heterogeneity can start with both naming and confounding conflicts For instance, a “river” and a “body of water” have two different realizations with respect to the scale and context of user perceptions A location-based query such as “find the intersections of a river and a highway in Ohio,” would need a semantic translator to identify the spatial object that cor-responds to the concepts referred by “river” and

“highway” The spatial agent needs to be able

to characterize the semantic content of spatial data and model The representation of “river” or

“body of water” depends on the context of the requesting agent Depending on the context, the geometric representations of “river” and “body

of water” could be also different; the resulting intersection of a “river” and a “highway” could

be a point feature in one representation while it could be a polygon feature in another

An agent’s access to data layers of spatial bases is dependent on the semantic interpretation and corresponding topological definition where spatial relationships such as adjacency relation-

data-ships can be reflected in a topological model A

topological model of a highway can be a graph

theoretic model characterized by its “from” and

“to” nodes as well as by its left and right polygon

objects An agent can interpret the semantics of

a textual query that includes “longest highway in

Ohio” in many ways since the agent would need

to understand the definition implicit in

topologi-cal containment of part and whole relationship

Examples of interpretation of “in Ohio” may

include or exclude highways that run along the

state border, that are wholly contained within the

state, or those crossing state boundaries Thus,

different interpretations will result in different

access and processing requirements to different

layers and spatial joins There is no standard or

common formal convention for the description and

categorization of semantic content in the query

which can be used for the global interoperability of

spatial agents In addition, the lack of

context-de-pendent data views can lead to monolithic systems

that do not promote interoperability For example,

it will be difficult for a local government or

busi-ness service provider to be able to update its land

records while serving them to other organizations

without transforming geometric and topological

semantics to make them compatible to the client’s

structure A state highway company would have

to maintain and reflect the changes made by local

governments for its own data view to be able to

use that local government’s base map directly At

present, there is no accepted convention, which

can be used to facilitate global interoperability

of spatial objects

o nt o Logy for seMAnt Ic

Interoper AbILIty

The objective of semantic interoperability is to

be able to attach meaning to information entities

or services and thereby draw inferences from the

semantic annotation In spatial semantic

interop-erability, the integration goes beyond the process

of data translation or the conversion of geometric

primitives The semantic interoperability ally boils down to the problem of the identification

eventu-of semantically similar objects that belong to different databases and the resolution of their se-mantic differences (Kashyap & Sheth, 1996) The use of an ontology (Guarino & Giaretta, 1995) as

a framework for defining similarity among objects has the benefit of a formal definition for concepts

in different metadata, a definition that could be used to define axioms for semantic translation between ontologies The term “ontology” has its root in the philosophical literature as the study of being In the domain of information systems and

AI, ontology has a somewhat different connotation

as an “explicit specification of a tion” (Gruber, 1993; Farquhar, Fikes, & Rice, 1996) and provides a more pragmatic definition: Ontologies are explicit specifications of domain conceptualization names and describe the entities that may exist in that domain and relationships among those entities In other words, the tacit and implicit knowledge hidden in a particular domain

conceptualiza-is explicitly conceptualized in ontology (Guarino, 1997) Ontology is considered as a logical theory accounting for the intended meaning of a formal vocabulary while conceptualizations are the formal structures of reality as perceived and or-ganized by an agent In spatial ontology, although agents may have a shared vocabulary capable of establishing relationships, or mapping between corresponding instances, the conceptualization

of space as “object-based” and “field-based” may

be still implicit among agents

Although spatial ontology is an established concept, and is capable of providing a naturalistic representation of spatial objects, in the sense of

“naive geography” (Egenhofer & Mark, 1995), it

is still a complex specification to be realized tial semantic interoperability goes beyond simple data translation or the conversion of geometric primitives based on a-priori ontology Classical ontology, offering limited expression of spatial relations to simulate spatial processes, can be used to express spatially complex phenomena,

Spa-is not well understood Specifically, such