|
GIS IMPLEMENTATION
TYPES
OF GIS IMPLEMENTATION
DESIGNING
YOUR GIS
Getting
Started
This excerpt [Guptill, Stephen C,
ed., 1988. A Process for Evaluating Geographic Information Systems.
U.S. Geological Survey Open-File Report 88-105.
Pages 11-22; 39-44.] was chosen to help you design a GIS that
specifically meets your individual goals. With this analysis, you can
determine your needs and what options are available before beginning this
process. It is suggested that you read and complete a User Requirements Analysis
(URA) before designing your GIS.
Chapter 3. User
Requirements Analysis (URA)
3.1
Overview
Geographic
information systems are successful when they comprehensively and consistently
meet the needs of users. Development of a successful GIS depends on well-defined
user requirements. A user requirements analysis (URA) is a detailed study of the
needs of potential system users. The URA should result in a clear statement of
end-product characteristics, required production rates, estimated data
volumes, and cost-benefit rationale.
In
performing a URA, much emphasis is necessarily placed on the detailed review of
the development and application of map-type products presently being required by
users. However, throughout the process, the URA manager should question the need
for traditional products and note GIS capabilities to meet the same requirements
with entirely new products and types of information and services in varied
formats. Users of hardcopy products may be limited in their present ability to
analyze and apply spatial information by the nature of the media in which they
deal. The URA manager may need to advise these users on the possibilities for
improving current capabilities of their organizations through GIS application,
not merely using the GIS to perform or mimic traditional functions.
3.2
Who Should Perform the
URA?
The
first problem faced by any organization considering the implementation of a GIS
is to determine who should perform the URA. URA’s for successful GIS
installations have been performed by in-house staff, contractor staff, or
through a combination of both approaches. There are many valid reasons for
opting for any given approach, however, the desired result is the same, to
develop a comprehensive assessment of the analytical capabilities and products
required by potential users of the GIS. The requirements of the users can then
be matched with system capability to determine optimal configurations for the
organization’s GIS procurement.
In-house
staff inherently have a greater understanding of the tasks which are to be
considered
for automation through GIS technology. This unique knowledge may justify
training staff members in GIS technology and URA techniques so that the URA may
be performed in-house. In cases where existing staff members have expertise in
GIS’s, there may be little reason to consider bringing in outside assistance.
When
staff time or skills are not available, or when new programs and concepts are
being proposed that staff is not experienced with, outside resources may be
required to perform the URA. Assistance may also be available from resources
within the parent agency of the organization considering the GIS implementation.
Assistance in developing Requests for Proposals (RFP’s) for URA services may
similarly be available within the agency.
The
important element in determining who should perform the URA is assuring that the
provider of the service has a thorough understanding of both GIS technology and
the operation of the organization. When an outside organization is brought in to
perform
the URA, it is the responsibility of the government technical representative
(GTR)
for the URA services procurement to assure that the contractor fully understands
the organization’s products, services, mission, and needs.
Possible
conflicts-of-interest should also be considered before a final determination is
made as to who should perform the URA. Organizations and individuals may, in
some instances, have a vested interest in certain hardware or software types,
and may be inclined, whether intentionally or unintentionally, to bias results
of the URA toward particular systems. The objective of the URA is to identify
the needs of an organization and then to select the system that best fills those
needs, if such a system exists. All reasonable effort must be made to assure
this goal is realized, including assessing possible conflicts-of-interests, or
biases, on the part of persons or organizations that potentially could perform
the URA.
3.3
Identification of Users
Once
resources have been identified for performing the URA, an initial step is to
identity the users of the proposed GIS. Depending on the nature of the
organization and its products, there may be several types of user groups to
consider. A user, in general, can be defined as a person who uses the system for
production, or who works with the products developed by the system. In some
instances, a single user will perform both of these functions, such as field
crews that both rely on the mapping produced by a system and also use the system
to update and revise their spatial data sets to produce new mapping. Another
category of user is the potential user, someone who cannot use the present
system because of some constraint, but could become a user of the system if it
were converted to a GIS.
The
system users are those who will actually interface with the hardware and
software to create the products of the organization. The function of a GIS for
these personnel is to replace or augment traditional cartographic, geographic,
and photogrammetric and related techniques, and to provide for more efficient
creation of products analogous to those developed in the hardcopy environment.
The system users may even apply the GIS technology for producing hardcopy maps
as a final product: the advantages of replicating traditional mapping through
the efficiencies of GIS technology being more important than the development of
new, digital products. More often the system users will be applying GIS
technology to provide new products and services. Although the system users may
only be developing a spatial data set for creating maps, by providing it in a
digital form for GIS application, they allow the traditional information to be
used in new ways.
The
end-product users employ the data created by a system, and the system-users, for the needs of their applications, programs, and
organizations. These users may be separated from the hardware and software that
is used in the GIS environment to create the products. If the end-product users
rely on hardcopy data, meeting their needs may only require the continuation of
the present hardcopy format data provided by the system. However, if the
end-product users desire to interact with the GIS system and use digital data
as opposed to hardcopy data, or will want new hardcopy products specialized for
specific applications, the GIS system to be implemented must be designed to
address these factors.
Factors
of concern when end-product users expect to work in the digital environment
include hardware, software, data translation, end-product user systems and end-product user applications. If the end-product user desires to interact with
specific hardware and software, the URA manager must determine the means for
accomplishing this objective. The end-product user may come to the site of the
GIS system, may use some type of computer network to tap system capabilities
from a remote terminal, or may send electronic files and programs to the GIS
site for execution by the system operators.
The
URA manager should determine whether the end-product users have the required
training and equipment to perform their desired applications on the GIS system.
When such training
or equipment is
lacking, the end-product users should be
so advised and planning ensued with the cooperation of the URA manager to
rectify these difficulties, or determine a more realistic approach for access to
the system by the end-product users in question.
Whenever
end-product users expect to take digital products from the GIS being implemented
and transport this data to their own systems for their applications, the URA
manager will need to determine the compatibility of the two systems. The URA
manager must answer not only the question of whether digital data can be
transported between the two systems, but also if the data can be transported
without loss of information. In some cases a software translator will be able
to reformat digital data from the implemented GIS to the end-product user
system, but a translator to perform the reverse operation may not exist, or may
not meet requirements. When end-product users expect to modify data sets and
return them for inclusion in the master GIS database, translators that will
operate effectively in both directions are needed.
End-product
user systems should be reviewed by the URA manager. In addition to issues
related to networking capabilities of computer systems and data translators,
there are other conditions that impact the relationship of the implemented GIS
with such related systems. Does the related system have storage capabilities to
manage files from the implemented GIS? Are the input/output capabilities
sufficient? Can it perform the analytical tasks required by end user
applications?
Ultimately,
whether the products from the implemented GIS will serve the needs of the
end-product user will depend on the user’s application. The URA manager must
clearly understand
the information and analytical needs of all potential users, and these needs
must be translated into GIS capabilities.
The
potential users of the system within the immediate organization performing the
URA usually can be identified readily by managers within the organization. These
may be both system users and end-product users. Organizational plans,
distribution lists, service contracts, and similar documentation can serve as a
guide for identifying all users of the data produced by the system proposed for
conversion to GIS technology.
Identifying
possible new users of data produced by the organization will be more complex.
These potential users will be characterized by data needs similar to those of
present users. However, the missions of potential users may be entirely
different from those of present system users and, therefore, their identities
not readily apparent. Some potential users will be organizations or personnel
who cannot presently apply products from the existing system because of some
constraint, such as map products at inappropriate scales, who could use products
that the GIS will be capable of developing. Other users may be those who could
use the products of the present system, but because of differing organizational
missions, their application was not recognized.
Identifying
these users will require researching the organization’s structure and missions
for similarities and identifying those that will be served by the proposed
system. Interviews with different offices can be used to form preliminary
impressions of the potential for new users. Although the identification of
potential users is a difficult task, efforts in this direction should not be
minimized. A GIS will provide the greatest benefits when utilized by the widest
range of users. Coordination of GIS capabilities with possible users prior to
implementation will be far cheaper then attempting to retrofit a system once it
is in place.
3.4
Definition of Required Products
The
purpose of a GIS is to create products of value to users. Before implementing a
GIS the products required by users must first be defined. Examples of products
required can be obtained directly from the present and potential users of the
organization’s data. These sample data sets should include both products that
the users are presently applying and also concepts and designs for desirable
products that could be created by a GIS. Through an analysis of these products
the URA manager can form some initial impressions of types of information, and
its media, format, and accuracy that the GIS will be expected to develop.
These
materials should be reviewed with the users to determine what aspects of the
information
content are required for user applications. Many general-purpose maps will
contain information that is not needed by some users. For example, the location
of power lines on a topographic map may not be important to a hydrologist
performing an analysis of a large watershed. Conversely, some information that
is not present on the current product may be required by the user and must be
created, or added to the existing product before the user can perform his
application. An example is the hydrologist performing the watershed study who
must first annotate the topographic base map of the area to show the watershed
boundaries.
In
addition to the information content of end-products, the URA manager should
apprise himself of user accuracy needs. In assessing accuracy needs a number of
sources may be consulted. The user may have published standards for the
application of specific types of information for specific purposes. Models may
have known accuracy limits that can be considered a upper bound for the accuracy
of required input data sets that might be derived from a GIS. Finally, the
present products used in the application can be reviewed to determine, if
possible, their accuracy.
The
preferred media of the user, scales at which map information will be required,
the format, including any color requirements, should all be noted at this stage.
The result of this review will be a preliminary identification of the types of
information required by users and the assessment of the needed media, accuracy,
and format.
Products
from a GIS can be broken down into several types, analogous to the breakdown
of users. Personnel involved in the operation of a GIS may use several products
at intermediate stages in the production process. Such products may be
quality-control check plots to assess digitizing accuracy and softcopy images to
preview products before hardcopy output. The final product normally developed by
a GIS is referred to as the end-product, which is applied by the end-product
user. Development of end-products is the major function of GIS in a map/spatial
information production organization. Such end-products and intermediate-products
can initially be defined in the terms described above.
GIS’s
that are being implemented to support analytical organizations that have no
standardized products and that are not concerned with the wide publication and
distribution of their data and information, pose a more difficult problem in
assessing product characteristics. These end-products may be reports, supported
by some graphics, that provide textual and tabular results of GIS analyses or
they may be more standard hardcopy maps. In some instances the organizations may
only be interested in creating digital products. These will require the
identification of digital data standards and digital data formats. Review of
past products and analyses performed by the organization and the mission and
goals of the organization will assist the URA manager in understanding the types
of products that a GIS must produce to support the organization.
3.5
Evaluation of Work Flow
Detailed
information on the current system, whether manual or automated, should be
gathered by the URA manager. This information should be gathered by interviewing
the personnel involved with the existing system, such as the managers,
professionals, and technicians. The interviews should be supplemented by
observation of their work, noting significant backlogs if any, and the
applications of various data sets that are made by the staff. Particular care
must be taken to identity the specific types of data used, including the data
topology, format, media, representation, and accuracy. At the same time
documentation concerning the costs of operating the existing system should be
developed. The results of the review of the existing system will form a base
line for comparison with any
proposed automated system and will serve as the basis for cost/benefit analysis.
Data
types can be classed by discipline. Within disciplines data can be further
sub-classed to provide an adequate description of data types that the proposed
GIS must manage and analyze. Each type of data will topologically represent
point, line, or area information, or in some instances may be used in different
spatial contexts for varying applications. For example, a stream gage record
represents point information, a water course represents line information, and a
river basin represents areal information. However, in a different application,
the user may be creating estimates of hydrologic parameters based on gage
records for application to watershed areas. In this instance the hydrologist has
applied point information to an area.
The
format of each type of data (tables, graphs, maps) should be noted by the URA
manager. Users should also be asked if the present format is most functional, or
could it be improved by conversion to an alternative representation that might
be more germane to the organization’s applications.
The
media in which data are used by an organization’s staff to accomplish its
missions should be reviewed by the URA manager to assure that a complete picture
of the data requirements is gained. As an example, soil maps are often prepared
on aerial photographs and the resulting compilation published. The user of the
soils map is provided, through the medium of the aerial photograph, with a great
deal more information than the areal extent and nature of soil types. Land cover
and land use data, as well as planimetric data are contained in the aerial
photograph. The URA manager may find that the users of the soil map not only
require the thematic soils data, but also rely on the ancillary data presented
in the aerial photograph to perform their functions. Identification of these
ancillary data for inclusion in the GIS data base should be made at an early
stage in the URA to facilitate planning for data base requirements.
Another
issue to be considered during the review of data used within the existing system
is accuracy. The transfer of these data to an automated environment should not
degrade the ability of system users to work with data that meet the accuracy
standards required by their mission. User data accuracy requirements may impact
GIS data capture design, coordinate system and map projection software
selection, and data output elements.
Data
can be symbolized many ways without changing information content. However, some
users may prefer drawings of double-line roads to single-line roads. Symbols for
representing point facilities, such as fire stations or hospitals, may have
widely recognized standards that should be adhered to and incorporated within
the GIS. The issue of graphic presentation is subjective. The requirements are
dictated by the user. The GIS system
most likely will allow
wide
flexibility in the presentation of information on products, and users may be
shown examples so that they will become familiar with the GIS potential for
developing data presentations that are tailored to their needs.
Often,
the staff currently producing a geographic product will not be the source of the
information used to develop the product, nor will they be the users of the
product. In these cases it will not be enough to simply review the
organization’s operations. Interviews and observations outside the
production organization among the suppliers of data
and the product users will be required. Information on data accuracy and the
potential for directly acquiring a digital product for loading into the proposed
GIS can be ascertained from the data suppliers. Product users can supply
information concerning the adequacy of present products, desires for alternative
products, and the minimum product that must be provided by the proposed GIS.
As
this stage the URA manager should have a complete picture of the data inputs,
production processes, data outputs, users, and costs of the existing system.
This information can be conveniently displayed as flow charts to provide a
system model for reference. Within the model information types, directions of
flow, production stages, and user interface can be depicted.
3.6
Data Base Development
The
data requirements of the users of the existing system will become the data
elements included within the data base supporting the proposed GIS. Often the
greatest cost associated with the development and implementation of a GIS will
be the effort to create a digital data base. Justification for applying GIS
technology to some application may not be sufficient to support the high cost of
collecting the required digital data. The URA manager will normally find that
there are needs that GIS technology can readily meet, but that these needs can
only be met if a complex digital data set is extant.
The
review of existing digital data sets for incorporation within the proposed GIS
is a first step towards data base development. The cost to create the required
data set may be mitigated if the full burden of data capture can be avoided.
Contact with such organizations as the U.S. Geological Survey’s National
Cartographic Information Center (that maintains a catalog of digital data bases
available from other agencies) and literature searches are a means to develop
information on available digital spatial data bases.
The
ability of existing suppliers of base data to provide their data in digital form
should also be investigated. It may be possible to have some forms of data
initially collected in a computer-compatible format and avoid later data
conversion. Examples are the use of computer-readable forms by field crews and
the use of digital remote sensing data for some types of land use/land cover
analysis missions.
The
final structure of the information, as it will be required by the users, should
be assessed to determine whether it will impact the means by which data are
captured. A map may be scanned at a large pixel size and provide adequate data
to users if the requirement is simply to perform grid-cell manipulations with
relatively crude grid dimensions. However, users requiring more precise
information may find such a data base unusable. More applications may be found
for data sets that are extremely detailed, but the costs of collecting, storing,
and manipulating these data must be weighed.
The
ability of a GIS to share information with other systems and to receive
information from other systems must also be considered. Networked computer
systems with distributed data bases provide the means for the power of an
individual GIS data base to be increased radically. The capability for data
sharing depends on the digital format of GIS data base elements. Consideration
should be given to the exchangeability of the data
format(s) used by the GIS’s proposed for the automation effort. The transport
of digital data between GIS without significant loss of information is essential
to maximizing use of spatial data bases.
3.7
User Applications
The
application of data within the GIS by users may require reformatting of data
structures and topology, interfacing with analytical models, and complex data
base management capabilities. Depending on the user needs, data may be handled
as raster or vector data. Each data format has advantages and drawbacks. The
applicability of a given data format will depend on the needs of the user. The
creation of map graphics and the analysis of data from a cartographic
perspective often requires vector data. Many environmental models are not
capable of managing vector data and operate based on raster data input, The
ability of the GIS to manage raster and/or vector data may be required for some
user groups.
Models
applied to the data base may be incorporated within the GIS or may be outside
the specific computing environment of the GIS. In both instances the ability of
data to be loaded into models in an automated fashion and for the results of
modeling to be portrayed by the GIS are important to users.
3.8
Refinement of GIS Product Characteristics
The
initial definition of required products and the evaluation of the current system
provide a description of products currently being provided to users. However, a
thoughtful reassessment of the characteristics of those products should be
performed before they are used as the criteria for selecting a GIS. The final
product definitions should reflect the flexibility of the GIS technology in
generating products, meeting the needs of user groups, and incorporating the
information gleaned during preceding activities.
In
many cases products from the existing system have been developed to serve a
large number of multidisciplinary users. As such, these products may contain a
large amount of information irrelevant to the applications of an individual
user. Or, an individual user may find a need to reformat, rescale, or extract
information from the existing product in some manner before these data can be
applied to the user’s application. In defining GIS products, flexible output
from a GIS should be evaluated for providing tailored products specific to user
applications.
Issues
relating to the formats, media, representations, and accuracy of data required
by the user must be carried over into the consideration of GIS end-product
characteristics. GIS output capabilities should be designed to provide the
product that the users require to perform their missions.
In
considering these issues, a differentiation must be made between a GIS that is
used to create a hardcopy product (perhaps simply emulating the existing system)
and one that is meant to support
analytical studies. When automation of a mass hardcopy production system is
considered, the desired GIS output may simply be a close facsimile of the
current product. Analytical procedures, even those within a single discipline
and performed within a single office, may differ drastically in required
end-products. Often tabular data are output from GIS-based analysis, while, in
other instances high-quality graphics will be required. Frequently a combination
of data presentations will be required. The URA manager must be aware of the
full range of desired end-product presentations.
Data
from GIS may be output to hardcopy or to an electronic media. GIS products may
be directly input to models, incorporated within digital data sets, or stored on
computer-compatible media. Users requiring these GIS end-products may have needs
for digital data formats and information types. In the design of the GIS the
input formats and data requirements of user computer models should be
reviewed. The objective of this effort is to provide an automated link between
user analysis of data and the GIS. An example is the modeling of river floods,
in which topographic data are used to delineate the extent of flooding. A
Digital Elevation Model (DEM) created by a GIS can serve as the base information
for this analysis, provided the DEM data can be output from the GIS directly
into the river model. Without such a linkage even though the DEM is an automated
product, the user may be forced to restructure and code the topographic data
into the flood analysis model manually.
A
definition of GIS product characteristics should result in a clear statement of:
•
Presentation media, Accuracy,
•
Information representation,
•
Information format,
•
Information content,
•
Digital product types and structures,
•
Analytical model interfaces.
3.9
Production Rates
Production
rates for the GIS must be specified so that the system can be designed meet
organizational objectives. If the GIS is intended to develop a well-defined
product for mass production the production rate is simply that stated for
the organization. The production rate in terms of digital data
can be calculated from an estimate of the digital data volume of a sample
of products.
For
a GIS intended to support analytical needs, the production rate is difficult to
quantity. Studies that will rely on a GIS to analyze spatial data are not
normally defined in terms of GIS products required to support the study. The
production of these analytical products of GIS is limited by the ability of the
system to retrieve, analyze, and display information from the data base. The
number of users expected to simultaneously be accessing a system and their data
processing requirements must be estimated to determine
GIS configurations capable of providing the level of user analytical support desired
by the organization.
Estimates
of required analytical products can be made by reviewing past studies by the
organization to estimate the number of final and draft map images, graphics,
statistical analyses, and other GIS products required. Interviews with staff can
identity additional products, which might have been produced if GIS capabilities
were available. Based on this past experience, the GIS production rates for
analytical studies can be extrapolated.
Hardcopy
production rates have four primary constraints:
•
Availability of digital data,
•
Availability of trained personnel,
•
GIS ability to process data,
•
Output device speed.
When
considering production rates, the types of products must also be identified. If
the product is intended to be a camera-ready image for mass reproduction and
distribution, high-quality output devices are required. Multicolored and
shaded outputs require specialized devices. The times to create a hardcopy
product from the digital data will vary depending on output device speed. Output
devices must be matched to end-product requirements and volumes.
The
number, types, and purposes of products required from the GIS must be quantified
before actual system configuration. When requirements have been identified they
can be compared with system abilities to produce similar products within
specified time limits.
3.10
Estimated Data Volumes
The number
of data capture, output, and storage devices, as well as the required sizes and
speeds of the CPU’s should be specified, in part, around the estimated volume
of digital data that will be handled by the GIS. Digital data will be input to
the system through either some data capture technique, or read directly from an
existing file. Aside from the density of graphic information to be digitized,
the volume of digital data will also be related to the data structure, format,
and level of detail captured. In some instances, particularly when raster data
are stored, significant compression of files can be achieved. When vector data
are captured, many points that are selected by the digitizer, or that result
from a scanning process, can be deleted without impacting the accuracy of the
line geometry.
As
a first step towards estimating data volumes, the efficiency of a format for
storing data relative to user requirements and applications, and the means for compressing these files should be considered. Data base size
should be minimized, and data structures simplified to the extent possible
without creating negative impacts on users.
Data
volumes can be estimated by having representative samples of data digitized in
the desired format and structure. The sample can be used to estimate the data
volume for the entire data base.
Data
can be classified by their function within the GIS, and data volumes can be
estimated for each type of data. Some data sets may not be utilized often and
can be stored offline. Similarly decisions can be made as to the storage
of data online in mass storage and the priority for retrieval of these data.
In
designing the overall GIS system, typical user activity on the system,
encompassing data capture, analysis, and output needs, should be estimated.
These estimates can be used to assure that the system configuration will support
user demands without undue hindrance from system access delays.
3.11
Cost/Benefit Analysis
A
cost/benefit analysis attempts to determine the costs of implementing a GIS and
to quantity the benefits that GIS implementation would provide above the present
system.
The
baseline for comparison is the cost of operating the present system.
Organizational budgets and records are the primary source of this information.
Cost factors, such as personnel salaries, hardware, maintenance,
overhead, and supplies, should be readily quantifiable.
Intangible
benefits from automation, such as improved response time and more flexible
output, will require comparison to some baseline. Existing system response time
records, products, and user interviews should be considered in generating the
baseline.
The
costs associated with the GIS should be calculated for comparison. These costs
include all purchase and maintenance costs for the system and should include the
economics of the system life cycle. Generally, information on these costs is
easily quantified, once a system or class of systems has been identified, from
educated analysis of vendor-supplied information.
Other
costs that must be considered are the costs for staff training, including lost
production while the staff becomes proficient on the new system. A change in
staffing also may be required to reflect the operating needs of the new system.
Perhaps
the most critical cost element is the creation of the required digital data
sets. This cost may range from 10 to 1,000 times the hardware and software costs
of the GIS. The GIS will not succeed without sufficient digital data. The cost
estimate for implementation must reflect
adequate data capture to support user needs.
Intangible
benefits, such as those associated with quicker response time or more product
flexibility, are difficult to quantity. Yet, many such benefits are the
strongest arguments for having a GIS. In each GIS cost/benefit analysis, the URA
manager should attempt to define intangible benefits and associate a monetary
value. The total cost/benefit results can be compared directly to the baseline
of the current system to view the economic
viability of implementing the GIS.
3.12
URA Report
The
URA manager should prepare a report for the organization management. This report
should clearly, and in detail, identity the following:
•
The operation, users, and data requirements of the existing system,
•
The potential users of a GIS,
•
Products required by users, digital and hardcopy,
•
Data volumes and production rates the GIS will be required to meet,
•
The data base required to support GIS implementation,
•
A cost/benefit analysis.
The
report should provide all the details required to understand the spatial data
that users require and the analytical capabilities needed to automate the
existing system.
Back
to top
-
Workstation
GIS - A
GIS operated in either a stand-alone environment
-
Enterprise-wide
GIS – A GIS
in which a series of PCs are connected by a network.; data resources
-
Web
distributed spatial data servers
-
Hybrid
GIS
Back
to THEORETICAL CONCEPTS IN GIS
Back
to TRAINING COURSE MAIN PAGE
| 