An approach for live data exchanges between ESRI ArcGIS and ...

DEVELOPMENT OF A COST-EFFECTIVE, USER-FRIENDLY ADMINISTRATIVE AND DATA MANAGEMENT SYSTEM FOR NEIGHBORHOOD ENVIRONMENTAL SURVEYS

Hua Zheng Health Geography Lab, Preventive and Behavioral Medicine, University of Massachusetts Medical School, Worcester, MA 01655, USA Email: [email protected]  

Christopher C. Duncan GISmatters, Amherst, MA 01002, USA Email: [email protected] 

Jennifer L. Kelsey

Health Geography Lab, Preventive and Behavioral Medicine, University of Massachusetts Medical School, Worcester, MA 01655, USA Email: [email protected]  

Wenjun Li*

Health Geography Lab, Preventive and Behavioral Medicine, University of Massachusetts Medical School, Worcester, MA 01655, USA Email: [email protected] 

*Corresponding author

- 1 -

ABSTRACT This paper introduces a user-friendly and cost-effective administrative and data

management system for neighborhood environmental surveys, using the neighborhood

walking environment as an example. The system uses a desktop geographic information

system application, ESRI ArcGIS, to manage geographically referenced information on

the environment in the study area, while concurrently employing the popular Windows

desktop database product, Microsoft Access, to administer field survey activities and

manage data collection. The integration of these two applications lets users with little

Geographic Information Systems expertise seamlessly and intuitively manage and

visualize the entire survey process in both map and tabular formats, including survey

planning, generating field assignments and data collection forms, processing collected

data, and tracking survey progress. The result is greater efficiency, improved data quality,

and significant cost-reduction in field survey administration and data management. This

paper describes the successful application of the integrated system to the assessment of

the design features and maintenance conditions of a neighborhood walking environment.

Keywords: Neighborhood Environmental Survey, MS Access Database, Geographic

Information System, Cost-Effectiveness, Data Management System, Project

Administration

- 2 -

BACKGROUND

This article describes a user-friendly neighborhood survey management system that

combines the familiarity and ease-of-use of form-oriented database software with the

visual and analytical advantages of geographical information systems (GIS) to provide a

seamless framework for survey planning, administration, tracking, visualization, and

analysis. Key benefits of this approach, which automates data exchange, manipulation,

and presentation between the database and GIS components, include: more efficient

planning and deployment of field assignments; lower skill-level requirements for data

entry, manipulation, and processing; improved data quality control; and more intuitive

and effective map-based tracking and analysis of survey results. In short, the system

provides better results with less effort at lower cost. We developed the system to support

a particular field survey of neighborhood environment, which we will use to illustrate its

features and benefits, but the approach can be readily adapted to other neighborhood

studies.

Neighborhood environmental policy can be a significant contributor to improvement of

public health. For example, living in a walkable neighborhood is important in promoting

physical exercise and general health (Nagel et al. 2008, Pikora et al. 2006, Li et al. 2005).

Good data on the neighborhood environment are fundamental to research into the

influences of that environment on the health of the residents and on related public policy

making. However, collecting and processing a large volume of survey data on

neighborhood environments in a reliable and cost-effective manner involves several

challenges, including the need to relate survey data to real-world features and locations to

- 3 -

model and analyze relationships between neighborhood environment and residents'

behaviors and health, and the problem of finding and funding staff with adequate skills

and experience in GIS software and methods.

Our research project, the Neighborhood Risk Factors for Falls in the Elderly Study,

supported by the National Institute on Aging, examines how the neighborhood walking

environment affects the physical activity and the risk for falling outdoors in a prospective

cohort of older men and women living in the Boston MA area. It is ancillary to the

population-based MOBILIZE Boston Study, which is following a cohort of 765

community-dwelling men and women aged 70 years and older. The MOBILIZE Boston

Study examines individual-level risk factors for falls, such as balance and gait, cognitive

function, medications, pain, musculoskeletal conditions, and foot disorders (Leveille et al.

2008). The ancillary study extends the scope of the MOBILIZE Boston Study by

integrating neighborhood environmental data with individual-level information on study

participants, and by examining how participant rates of outdoor falls vary according to

neighborhood characteristics. The walking environment is of particular interest in this

study. We developed a set of surveys, the Older Pedestrian Environment Survey (Figure

1), as a means to assess objectively the physical attributes of the walking environment in

the neighborhoods in which the study participants reside. The Older Pedestrian

Environment Survey assesses the design features (e.g., sidewalk availability, width of

sidewalk, type of pavement) and maintenance conditions (e.g., sidewalk surface condition

and cleanliness) of the walking systems during the study period.

- 4 -

Figure 1. Example of field survey forms. The parts of the Older Pedestrian Environment Survey are given, including Field Survey Tracking Form, Street and Sidewalk Features, and Sidewalk, Street and Road Conditions.

- 5 -

Sidewalk spatial data, collected and processed for the study, consist of two components,

that is, geographic reference data and field survey data. The geographic reference data

refer to information on 1) geographic units defined by the US Census (e.g., census tracts,

census block groups and census blocks), 2) municipalities defined by the Commonwealth

of Massachusetts (e.g., cities, subdivisions of large cities and towns), and 3) road

networks within the study area. The geographic reference data contain information on

these geographic units, such as their specific geographic coordinates and their topology

where points represent street/road intersections, lines represent street/road segments and

polygons represent census units and municipalities. The field survey data are collected by

field staff through on-site inspection of the design features and maintenance conditions of

street segments and intersections and linked to those geographic reference features to

enable map display and spatial analysis of the survey results.

We use Microsoft Access 2007 for form-based survey data display and management, and

ESRI ArcGIS 9.3 for map-based geographic data display and management. Microsoft

Access, the most popular Windows desktop database management system, has been

commonly used in public health research settings because of its wide availability,

relatively low cost and remarkable ease of use. We therefore chose it to handle the

collection and data-entry aspects of the survey for this study. ESRI ArcGIS is a widely

used desktop Geographic Information Systems (GIS) software package developed for

processing and analyzing spatial data. Its powerful map-based visual display capability

allows the depiction of complex spatial information in an intuitive and easy-to-

comprehend format (Clarke et al. 1996, Skop et al. 1999, Rushton 2003). However,

- 6 -

ArcGIS does not provide effective tools and functions to facilitate the collection and

management of survey-based data without extensive programming. Our experience from

previous studies also indicates that ArcGIS requires extensive personnel training before it

can be used easily and effectively by survey administrators and workers; among public

health research institutes, there are relatively few personnel with adequate training in

ArcGIS to be able to operate such an information management system.

Independently managing separate information systems requires time-consuming and

error-prone procedures to keep data synchronized (i.e., concurrent data updates across

two systems). There are various limitations of using separate information systems. First,

if any changes are made to the samples of geographic features in ArcGIS, such as adding

or dropping a sampled object, data management staff have to manually insert or delete

the associated geographic reference record in Access in order to ensure the consistency of

geographic reference data in ArcGIS and Access. Second, after entering daily survey

data into Access, data entry staff immediately have to update the status attributes of

surveyed units in ArcGIS in order to display the study progress in map format in ArcGIS.

Third, the staff have to switch frequently between ArcGIS and Access in order to review

both the geographic location of a sampled unit and its associated survey records. Fourth,

the survey records in Access are managed and displayed by alphanumeric codes in tables

or forms rather than visually in a geo-referenced format, making it difficult to locate

specific geographic units and to track study progress. Each of these tasks across the

separate information systems presents the risk of errors in identifying and modifying

features and records, which would compromise overall data consistency, completeness,

- 7 -

and accuracy. In summary, maintaining separate data entry and mapping systems creates

data exchange and data integrity problems that require continuous, careful and costly

efforts to address.

To eliminate these problems and enhance data management operations, we developed an

integrated data management system with live and automated data exchange modules that

allow ArcGIS and MS Access to interact continuously and seamlessly. The integrated

data management system is designed to: 1) achieve greater efficiency in data delivery by

minimizing or eliminating error-prone manual operations; 2) be cost-effective and easy to

use in field survey administration and data operation; 3) have reliable and fast processes;

and 4) minimize staff training requirements. Following these design goals, we developed

an approach that uses Access to generate field survey assignment and data collection

forms, process collected data, and track survey progress, and ArcGIS to concurrently

visualize and process spatial information on the geographic units under examination,

update study progress, and produce survey site maps.

METHODS Geographic Reference Data in ArcGIS

The ArcGIS database for the Older Pedestrian Environment Survey includes detailed

geographic reference information on road segments and intersections derived from the

local road network data obtained from Massachusetts Geographic Information Systems

(MassGIS), and US Census 2000 information on the census block, block group, tract, and

city/town of those particular segments or intersections. Three types of spatial units were

- 8 -

used in this system: census block groups, road/sidewalk segments, and road intersections.

For each type, an associated map layer was created in ArcGIS.

Within each block group, a random sample of non-contiguous road segments and their

connected intersections was selected for field observation using a sampling algorithm

written in Stata MP 9.2 (Stata Corp., College Station, TX). This algorithm approximates

uniform spatial distribution of selected segments and intersections and thus optimizes the

geographic representation of collected data. In total, 6,500 segments and 10,500

intersections were sampled from the 611 block groups in the study area. Site maps, at the

level of community and census block group, were produced as a guide for daily survey

assignment and tracking of field work. On the site maps, spatial units were color-coded to

reflect their status and status changes, such as sampled versus non-sampled units or

completed versus non-completed field survey and data entry. This is illustrated in Figure

2.

- 9 -

Figure 2. Example of block group-level survey site map. A block group with ID 250250805003 in the South End community of City of Boston was selected as an example to show its site map at the block group level, which is printed and used as a guide material for daily survey assignment and tracking of field work.

- 10 -

Sidewalk Survey Data in Access

The Access database is primarily used to store and process the data from the field survey

done by our staff. Since characteristics differ between road segments and intersections

from the environmental perspective, their survey questions were designed separately. For

example, the intersection surveys focus on the point-based corner and crosswalk elements,

unlike the line-based street and sidewalk attributes from the segment surveys. Therefore,

the four primary tables (hereafter referred to as sidewalk survey tables) were set up to

store information on the four categories of field survey data: segment features, segment

conditions, intersection features and intersection conditions, respectively. Additional

survey completion status fields, such as data entry completed or not, were also set up for

all survey records in the four tables. In order to keep track of the relationship between

geographic reference features (e.g., road line segments) and survey data (e.g., sidewalk

conditions), the sample segment and intersection lists were separately imported into two

additional Access tables that explicitly link the record identifier of each geographic

feature with the record identifier of the corresponding survey record for that feature (in

database parlance, these tables are often called "mapping tables", because they "map"

records from one table to records in another; to avoid confusion with our use of the terms

"mapping" and "maps" in their cartographic sense, however, we refer to these tables

hereafter as geographic reference tables). We sampled segments and their connected

intersections. For convenience, each intersection is uniquely identified by two

intersecting segments. The sampled segment connecting to an intersection is named

Intersecting Segment A, and the adjacent segment in clockwise order is defined as

Intersecting Segment B, such that segments and their connecting intersections can be

- 11 -

linked using segment as well as intersection identifiers. The data structure of the Access

database is shown in Figure 3. Some of the geographic location attributes are included in

the Access database as well, such as Block Group Number, to generate field survey

assignments and data collection forms used together with printed site maps for daily

survey deployment.

- 12 -

Figure 3. Access database structure. The boxes represent tables containing geographic and survey data; the black arrows on the right represent links between geographic and survey records.

- 13 -

Integration of ArcGIS and Access

The integration is achieved by means of code modules that concurrently manage both

ArcGIS and Access data and user-interface components. The modules allow the two

applications to communicate and exchange geographic reference data or sidewalk survey

data in either direction (Figure 4), with the simultaneous display of corresponding data in

Access forms and ArcGIS maps. With the integrated system, data updates in ArcGIS or

Access make immediate and automatic corresponding changes in the data tables of the

other application; this facilitates the study process and makes the field survey

administration and data management system more efficient and cost-effective.

Figure 4. Live data exchange modules. Four designed modules allow the ArcGIS and Access databases to communicate and exchange geographic reference data or sidewalk survey data in either direction.

- 14 -

The code modules automate four distinct types of interaction between Access and

ArcGIS: two modules operate within ArcGIS to handle user actions made in the map

display that require corresponding updates in the Access tables, one operates within

Access to handle user actions made in the forms that require corresponding updates of the

GIS data, and the other symbolizes and labels within ArcGIS survey-related features

delivered from Access. One ArcGIS module automatically updates geographic

information on sample spatial units in the Access geographic reference tables in response

to simple map-based sampling or non-sampling operations, keeping the sample data

synchronized between the two applications without manual transmission of data from

ArcGIS (Figure 4 - ). A second ArcGIS module opens a data entry form in Access for a

specific segment or intersection record by clicking on that feature in the map display

(Figure 4 - ), making it possible for less skilled workers to carry out data entry tasks,

therefore reducing operational costs. The module in Access provides complementary

functionality: the module lets users select and highlight in the ArcGIS map interface the

feature corresponding to the current record being viewed in Access (Figure 4 - ),

providing an easily understood display of the surveyed unit locations for data entry staff

and an intuitive way for data managers to monitor the field survey work. The fourth

module with joint functionality immediately displays on the map within ArcGIS the

survey results and status entered in Access (Figure 4 - ).

Approaches to Programming the Integration Functions

The integration functions have been implemented with the Visual Basic for Applications

(VBA) development environment in ArcGIS and Access, respectively, on a Windows XP

- 15 -

platform. VBA has been used for development of user interface and database operations

in ArcGIS (Patil et al. 2008, Marinoni 2004) and Access applications (Hanchette et al.

2005).

From the ArcGIS side, new user-interface (UI) controls have been created to execute the

interactive operations with Access. For convenience, the newly established controls are

arranged on the Tools bar. The operation of these controls is the same as any other button:

clicking a control and then clicking the map enables that selected control function.

Clicking on any of the controls starts up the Access application (if it is not already open)

and makes the appropriate function calls and data exchanges to carry out the requested

operation. The controls are feature-aware, meaning that clicking on different feature

types in ArcGIS, such as block groups (polygons), segments (lines) or intersections

(points), triggers appropriate processing in Access by using layer category filtering.

Specifically, the “Add Info” UI control is used to create a new sample of segment or

intersection and to insert its corresponding geographic data into the Access geographic

reference tables. After selecting the “Add Info” control, clicking on a non-sampled

segment adds a new set of segment geographic data into the Access segment-related

geographic reference table, and clicking on a non-sampled intersection adds a new set of

intersection geographic data into the other intersection-related geographic reference table.

Checks are made before every insert operation in the Access tables to ensure the feature

does not already exist in the current sample list. Conversely, the control named “Delete

Info” changes the unit status from sampled to non-sampled on the map and

simultaneously removes the existing information related to that particular unit stored in

- 16 -

the Access tables. The UI control named “Connect to Survey Info” navigates users

directly to the survey entry forms in Access. If a segment or an intersection is clicked in

ArcGIS, the corresponding segment or intersection record is automatically opened in an

Access form; if a blank place in a block group is clicked in ArcGIS, a complete list of

sampled segments in the block group appears in tabular form in Access, each row

represents a specific segment; by clicking on any place in a row, one can open the survey

form of that segment.

From the Access side, a dual-function button linked to the ArcGIS map, called “Zoom to

Feature”, is created for every segment and intersection unit. Clicking on the button

associated with a segment or an intersection survey record in Access launches ArcGIS (if

it is not already open) and the specific segment or intersection is selected on its respective

map layer, with the zoom extent centered on the feature.

In addition, join operations are created in ArcGIS to build a live link appending the

required fields of the Access sidewalk survey tables to the associated segment or

intersection attribute tables in ArcGIS. The joint operations are implemented using OLE

DB providers and ODBC drivers that are available in ArcGIS 9.3 and enable easy access

to external data in Access. OLE DB connections have been commonly used in GIS

studies to retrieve data from databases (Kim et al. 2002). With the join function, the

sample units as well as their survey data (survey results and survey status) can be

concurrently displayed on a map, categorized and marked using different symbols and

colors, which enhances the visual impact of the various spatial data in the study.

- 17 -

RESULTS

The new integrated data management system with live data exchange modules results in

significant improvement in several aspects, including a) greater efficiency, b) more user-

friendly functions, c) better reliability, and d) lower staff skill requirements.

The new integrated system is about five times more efficient than the older system

without any integration. To compare the efficiency, we used as a benchmark the time

needed to complete pre-field work preparation and data entry for a typical block group

with 10 segments and 20 intersections. With the non-integrated system, it took over 45

minutes for a well-trained research assistant with a bachelor’s degree in geography to

prepare survey forms and track/update the status of target segments; with the new system,

it took less than 8 minutes for the same person to complete the same amount of work.

With the non-integrated system, a trained research technician took about 3 hours to

complete the data entry, and an additional hour to verify the accuracy of data entry and to

update and synchronize the two databases. With the new system, data entry for a block

group was completed in less than 1 hour. Data quality check and synchronization of the

databases are automatically executed without any manual operations.

Through the user-friendly interface design, which provides simple and familiar button-

based "tools" to activate the automation code running in the background, the data

management staff can easily update and organize the study sample environmental units in

either of the two databases without worrying about data discrepancies between the two

- 18 -

databases, and the data entry staff can also easily operate and review geographic data and

survey records in a smooth and automated way, without frequent manual switches

between the two applications.

The integrated system is more reliable since it ensures that requested actions will be

carried out appropriately by using feature-specific processing. For instance, clicking on a

segment (a line), an intersection (a point) or a block group (a polygon) in ArcGIS with

the UI controls causes similar but feature-specific operations to be performed in the code

(Figure 5). Likewise, the “Zoom to Feature” button in Access responds appropriately to a

request for visual display of location whether the record is for a road segment or an

intersection (Figure 6).

Lower technical skill requirement and user-friendly interface design make system

operation simple and cost-effective. Since the UI controls hide the technical

programming from the user, the function operation and data management of the two

software applications, particularly those of ArcGIS, are much easier; little or no

experience or training in ArcGIS is needed for daily data entry and data organization

work.

- 19 -

 (a) Data entry location of a specific segment

 (b) Data entry location of a specific intersection

Figure 5. Response examples of the UI control “Connect to Survey Info” in ArcGIS. (a) Clicking the UI control “Connect to Survey Info” and then click on a segment with ID 54085 in ArcGIS launches Access and directly navigates to the data entry location of the selected segment. (b) Clicking the UI control “Connect to Survey Info” and then click on an intersection with ID A54085B52363 in ArcGIS launches Access and directly navigates to the data entry location of the selected intersection.

- 20 -

 (a) Geographic location of a specific segment

 (b) Geographic location of a specific intersection

Figure 6. Response examples of the button “Zoom to Feature” in Access. (a) Clicking the button “Zoom to Feature” for a segment with ID 54085 in Access launches ArcGIS and zooms in to the geographic location of the selected segment. (b) Clicking the button “Zoom to Feature” for an intersection with ID A54085B52363 in Access launches ArcGIS and zooms in to the geographic location of the selected intersection.

- 21 -

DISCUSSION

We have successfully designed and developed a data management system with a live data

exchange capability, by integrating the functions between ESRI ArcGIS 9.3 and MS

Access 2007. This system facilitates the data management operations of the

environmental data collected in this Neighborhood Study. The system provides an

efficient means to perform data operations across the two applications and shares

geographic reference data and sidewalk survey data instantly in both applications. First,

the seamless retrieval and visual display of survey data in ArcGIS and the quick zoom

function identifying the particular geographic location of a unit in Access improve

efficiency and provide users with an intuitive interface to control concurrent changes.

Second, the ArcGIS and Access code modules bind the previously unrelated processes

and ensure the integrity of the database structure of the combined system. Data

communication becomes more efficient with the tight integration of the two applications.

Third, the concurrent operations in both databases result in significant labor-saving and

cost-reduction in field survey administration. Finally, the new system does not require the

operators to have previous experience in the use of ArcGIS, which is an important factor

in public health workforce where finding a staff with ArcGIS training is not common.

To develop such a system, a person with intermediate level of programming skills in

database design is needed. We estimated that the cost for developing such a system is

around 700 working hours in academic settings without prior experience, with

approximately 160 hours spent on the development of GIS data layers and a sampling

plan, 200 hours on the construction of Access databases, and 340 hours on the integration

- 22 -

of the two databases. The cost may be substantially lower if the database design team has

prior experience. Considering the potential significant savings on labor costs for field

work administration and data processing and remarkable improvement in data quality,

our approach to integrating ArcGIS and Access data is highly cost-effective.

Nevertheless, our current implementation has several shortcomings. First, the integration

between ArcGIS and Access is only at the level of specific data operation functions; this

is considered "partial integration". Both applications still operate in two independent

software packages. If a computer with a single monitor is used, a user has to switch

windows manually between the two applications. To avoid this, it is necessary to use dual

monitors connected to the same desktop computer, or to use a single wide-screen monitor.

A better and more efficient method is to achieve a "full integration" though at greater

software development and maintenance, through which the functions available in Access

are completely integrated into ArcGIS, and user interface functions are created within

ArcGIS to allow users execute all Access data operations within ArcGIS. One promising

approach is to build an integration interface in ArcGIS for survey data entry, through a

complete back-end data exchange with the Access database. As a component of ArcGIS,

the survey data interface would offer more effective data retrieval and storage with

spatial data organization. Another potentially effective solution would involve the use of

a handheld device, such as, a PDA (Casademont et al. 2004, Huang et al. 2004, Lin et al.

2006, Cardoso et al. 2004), to input field survey data directly into ArcGIS and Access

through web services instead of traditional paper-based recording of field data.

- 23 -

CONCLUSION

In conclusion, a new efficient integrated survey administration and data management

system was developed through the implementation of seamless and live data exchange

between ESRI ArcGIS and Microsoft Access. This integrated system provides users with

an interactive interface in an intuitive, easily understood format when making and

viewing concurrent changes in the two application components, resulting in greater

efficiency and significant cost-reduction in field survey administration and spatial data

operations. The same approach could be applied to data management systems for the

surveys of park and recreational areas, neighborhood grocery stores, and other business

establishments. We encourage other public health researchers and practitioners to apply

this method when conducting community environmental surveys.

- 24 -

About the Authors

Hua Zheng is a Postdoctoral Research Associate in the Health Geography Lab, Division

of Preventive and Behavioral Medicine at the University of Massachusetts Medical

School (UMMS). Her research focuses on the use of information technology to improve

health service research and health care delivery.

Christopher C. Duncan is President of GISmatters, Inc. and Adjunct Professor in the

Department of Geosciences at the University of Massachusetts Amherst. His interests

include application of Geographic Information Systems, spatial analysis, systems

integration, and database development.

Jennifer L. Kelsey is Professor of Medicine in the Division of Preventive and Behavioral

Medicine at the UMMS. She has a long-standing interest in epidemiology and prevention

of falls and fractures.

Wenjun Li is Director of Health Geography Lab and Associate Professor of Medicine in

the Division of Preventive and Behavioral Medicine at the UMMS. His research focuses

on analyzing impact of neighborhood environment on health and health behaviors.

Acknowledgments

This research was supported by the National Institute on Aging (NIH/NIA

1R01AG028738). The authors thank Dr. Sylvie Puig for her editorial assistance, Ms.

Danielle Fontaine, Ms. Laura Mottola, Mr. Daniel Makridakis and Mr. Alex McManus

for their suggestions and testing the integrated system.

- 25 -

References

Cardoso J., A. Rocha and J.C. Lopes, 2004, M-GIS - Mobile and interoperable access to

geographic information. Electronic Government, Proceedings, 3183, 400-405.

Casademont J., E. Lopez-Aguilera, J. Paradells, A. Rojas, A. Calveras, F. Barcelo and J.

Cotrina, 2004, Wireless technology applied to GIS. Computers & Geosciences,

30(6), 671-682.

Clarke G.P. and R. Langley, 1996, A review of the potential of GIS and spatial modelling

for planning in the new education market. Environment and Planning C:

Government and Policy, 14(3), 301-323.

Hanchette C.L., D.A. Gibbs, A. Gilliam, K.J. Fogarty and M. Bruhn, 2005, A national,

geographic database of CDC-funded HIV prevention services: development

challenges and potential applications. International Journal of Health Geographics,

4, 28.

Huang B., T.F. Fwa and W.T. Chan, 2004, Pavement-distress data collection system

based on mobile geographic information system. Pavement Management,

Monitoring, Evaluation, and Data Storage, 1889, 54-62.

Kim D.H., M.S. Kim, 2002, Web GIS Service Component Based On Open Environment.

IEEE International Geoscience and Remote Sensing Symposium, 6, 3346-3348.

Leveille S., D. Kiel, R. Jones, A. Roman, M. Hannan, F. Sorond, H. Kang, E. Samelson,

M. Gagnon, M. Freeman and L. Lipsitz, 2008, The MOBILIZE Boston Study:

Design and methods of a prospective cohort study of novel risk factors for falls in

an older population. BMC Geriatrics, 8(1), 16.

- 26 -

- 27 -

Li F.Z., K.J. Fisher, A. Bauman, M.G. Ory, W. Chodzko-Zajko, P. Harmer, M. Bosworth

and M. Cleveland, 2005, Neighborhood influences on physical activity in middle-

aged and older adults: A multilevel perspective. Journal of Aging and Physical

Activity, 13(1), 87-114.

Lin C.C., M.J. Chiu, C.C. Hsiao, R.G. Lee and Y.S. Tsai, 2006, Wireless health care

service system for elderly with dementia. Ieee Transactions on Information

Technology in Biomedicine, 10(4), 696-704.

Marinoni O, 2004, Implementation of the analytical hierarchy process with VBA in

ArcGIS. Computers & Geosciences, 30(6), 637-646.

Nagel C.L., N.E. Carlson, M. Bosworth and Y.L. Michael, 2008, The relation between

neighborhood built environment and walking activity among older adults.

American Journal of Epidemiology, 168(4), 461-468.

Patil J.P., A. Sarangi1, O. P. Singh, A. K. Singh and T. Ahmad, 2008, Development of a

GIS Interface for Estimation of Runoff from Watersheds. Water Resources

Management, 22(9), 1221-1239.

Pikora T.J., B. Giles-Corti, M.W. Knuiman, F.C. Bull, K. Jamrozik and R.J. Donovan,

2006, Neighborhood environmental factors correlated with walking near home:

Using SPACES. Medicine and Science in Sports and Exercise, 38(4), 708-714.

Rushton G., 2003, Public health, GIS,and spatial analytic tools. Annual Review of Public

Health, 24, 43-56.

Skop E. and J.S. Schou, 1999, Modeling the effects of agricultural production. An

integrated economic and environmental analysis using farm account statistics and

GIS. Ecological Economics, 29(3), 427-442.