ABSTRACT: The County of Chester, Pennsylvania has been in the process of implementing a geographic information system (GIS) for the past seven years. The project initiated with the establishment of a project team consisting of county staff and a consultant company to mange the GIS implementation. The team was charged with the implementation of the GIS with development of some reengineering processes for departments involved with land records transactions. Two committees, Policy and Technical, consisting of representatives from various departments, were also formed to assist in the GIS implementation. The people involved and their approach to project management and decision making established commitment for project completion. This paper describes the project management strategies; accomplishments and lessons learned in implementing a countywide GIS.

Chester County Location Map

GIS Initiatives

In 1988, the County retained the services of an outside GIS consultant to assess the implementation of geographic information system (GIS) in the County. The needs assessment report consisted of an overall review of the departments and their needs for interdependencies. It identified the core departments that would benefit from implementing a GIS. That report was accepted but put aside. In late 1992, Chester County requested that the consultant update the needs assessment report to generate a Master Plan and prepare a cost estimates for the acquisition and implementation of a GIS to be reviewed and adopted by the County Commissioners. During the same time the Emergency Services (ES) department was upgrading its installation with computer-aided dispatching (CAD) and an E911 system. To support the CAD system they were in need of GIS data consisting of road centerlines with accurate street names, address ranges and addresses. ES pursued a separate vendor and contract for the Addressing Project, but coordinated with the Master Plan efforts.

The Master plan "for a land records system for Chester County" included a concept and the components of a proposed GIS system, and the tasks and schedule for implementation of the system. The master plan was based on the review of then current land records operations and an evaluation of requirements for land records. The concept addressed the requirements for the hardware/software configuration and database to support the necessary information and its management for the GIS. The concept was based on a central database management system and sharing of data and other system resources among county departments. The GIS was the key part of this concept to provide needed mapping and geographic analysis capabilities. A series of applications to support or perform various functions of participating departments were identified. The overall concept for implementation of GIS included a central database management system, continuous updating and maintenance of data, to provide necessary up-to-date data to all system users and applications. The project management and implementation strategies are covered in a later section of this paper.

Modernizing land records was defined as integrating the Recorder of Deeds processes and data with those of the Assessment Office. By the adoption of the Pennsylvania Uniform Parcel Identifier (UPI) Law, business process reengineering and computer automation, the County could streamline workflows, eliminate redundancies, store and retrieve data more efficiently, and provide better service to the public. Under the UPI Law, all recorded documents affecting the future transfer of land must have an UPI on them, and the UPI would be the legal equivalent of a legal description. Access to all records relating to any particular parcel can be made by the UPI assigned to that parcel, as well as the traditional keys of owner name or street address.

The Land Records System (LRS) is one of several enterprise-wide automation projects the County has undertaken in the past decade. The migration of many legacy applications running on a Unisys mainframe to a client/server architecture spawned many enterprise-wide infrastructure improvements and exponential growth of the network. Such projects as a county-wide reassessment, the acquisition of the Integrated Assessment System (CAMA), the E-911 system, the Courts system and the PeopleSoft Enterprise 2000 have been undertaken in an attempt to position Chester County as a leader in government services in the 21^st century. While the vision and scope of these projects are extraordinary, the challenges in completely realizing the benefits of these efforts are sometimes daunting. A modern Land Records System yields many benefits not only to Chester County, but to the municipalities, real estate and title professions, and the general public as well. Having a centralized database and appropriate automation to support county operations will revolutionize the way Chester County government does business.

Conceived in 1993, the initial LRS was designed, implemented and tested between 1994 and 1997. Two committees were formed to oversee the project: the Technical Committee and the Policy Committee. Working with these committees, the Department of Computing and Information Services (DCIS) and consultants undertook the responsibility of reengineering the Land Records process. This endeavor was composed of several significant tasks:

1. Establish the Business Rules, Procedures and Data Model
2. Develop Institutional Restructuring and Process Reengineering
3. Application Development – Providing the Tools for Implementation

These tasks were attempted over a period of seven years and purported to lay the foundation for a LRS. As projected, the LRS project succeeded in:

• Establishing a high-level systems architecture and a GIS Organizational Structure.
• Establishing a Registry Organization, the Bureau of Land Records (BLR), as the central and permanent repository of all County Tax Maps, and as the designator of an UPI.
• Acquiring the necessary hardware and software systems needed to support a state-of-the-art GIS, including a modern LRS.
• Performing the digital Tax Map Conversion and acquiring a planimetric Land Base, both of which form the foundation for the County’s digital map database.

These accomplishments have provided a framework for all future GIS and land records activities.

A brief history of GIS implementation by the County is presented in Table 2. The Table 2 lists major accomplishments and milestones by year. Figures 2 and 3 illustrate capital spending history and breakdown in the implementation of LRS during 1993-1999.

The County’s Department of Computing Information Services (DCIS), was charged with the implementation of an enterprise-wide GIS, consisting of a land records system, a relational database management system, ArcInfo and ArcView software, in addition to imaging and electronic document management software. Departments actively participating in the implementation GIS are listed in Table 3.

LIST OF PARTICIPATING DEPARTMENTS

The basic strategy to manage and perform the implementation of GIS was to set up a land record project team, consisting of staff from DCIS and personnel from a consulting firm.  In addition, two committees, a Policy and a Technical Committee were formed to assist in the implementation of the GIS.  The team was supervised by the County's Land Record Area Manager and was responsible for the implementation of the GIS.  The County initially did not have enough resources or people trained in GIS and, therefore relied on the consultants expertise in GIS field.  There were on-site consultants, who were working full-time on the project to assist the county in the implementation effort. The lack of GIS implementation experience put the county staff at disadvantage.  They had to rely on the consultants expertise.  While the LRS project has benefited from a team approach and the involvement of multiple departments, consultants and committees, it has also suffered traditional disadvantages, such as departmental inflexibility, lack of coordination, staff burnout and slow response to change. 

Policy Committee consists of executive level persons. Its role is to assist in establishing policies and priorities, to guide system development and its evolution, and assist in budget co-ordination. Membership for this Committee is drawn from key user departments (initially DCIS, Planning, Assessor, Recorder, Health and Emergency Services). Representatives of individual departments shall identify key GIS personnel within their department, define how to employ GIS technology, participate in and review conceptual database design, provide guidance in the development of internal procedures and data access policies.

Technical Committee consists of technical level persons. This group addresses technical issues and provides guidance to Land Records System core staff and users. Its role is to identify policy issues for Policy Committee consideration. Membership for this Committee is drawn from user departments (initially DCIS, Planning, Assessor, Recorder, Health and Emergency Services). Members shall, within their respective departments, be very familiar with the internal processes, be able to identify needs, users, training and design applications. They should also be intimate with the implemented technology so they can provide recommendations/requests for system enhancements, extensions or procurement of supplies.

This is the initial Land Records System (now the GIS Technical Services group) core group. It provides system management, database administration and day to day support to system users. Its responsibility is to manage all hardware and software components, perform on daily basis data and software backups, resolve problems, liaison with vendors, administer maintenance contracts, install upgrades, monitor system resource utilization, plan and acquire replacement/upgrade/enhancements, authorize system access and allocate system resources.

On the database side this group provides central administration of the LRS database, authorize database access privileges, maintain database directories, maintain and upgrade database design, ensure database integrity and implement quality assurance procedures.

On the development side this group designs, develops and maintains application programs, determine user requirements, provide consultation services for users and document user applications.

This group also identifies technical issues that are reviewed by the Technical Committee for guidance.

Over time the role of the DCIS Land Records Team has changed and responsibilities have been assigned to two groups. A GIS/BLR group is responsible for maintenance of parcel database and some application development. A GIS Technical Services Group is responsible for much of the hardware and software management activities.

It is essential that all GIS resources be managed carefully and continuously. Because of the number of departments and agencies involved, specific plans for managing, maintaining, and distributing data should be developed. Currently the GIS implementation is managed by two groups, as mentioned above.  The GIS Technical Services Group, who manages and administers the GIS hardware and software and database.  This Group has responsibility for providing GIS and Information Systems support to County staff. The County potentially has more than 100 users of the GIS. The GIS/BLR who develop and maintain GIS data, both graphic layers and nongraphic databases, establishes data development priorities, and oversee contractor effort for database and application development.  This Group has responsibility for maintaining one of the most important GIS resources, the parcel base.  The GIS Technical Services Group must ensure that it is published and accessible to its large user base.   

The GIS serves all County departments. As more and more GIS users can be served and brought on line, the benefits of full GIS implementation will begin to be realized. The GIS data resources are also extremely valuable assets of the County. These vital and dynamic resources must be carefully maintained in order to retain their value. Projects such as IAS (Integrated Assessment System) upgrade, which are underway will result in significant integration of data resources as well as updates of operating environments.

The GIS is managed and administered as an IT system under the direction of the County Chief Information Officer (CIO). As a unique system, however a specific core GIS Group was created within the DCIS structure. Significant GIS activity does occur at the department level. Departments who have developed GIS resources and expertise can rely on the DCIS for support and coordination. DCIS provides technical support and assists with the gradual improvement and centralization of enterprise-wide resources such as the fundamental GIS layers. Individual departments may still develop mission specific GIS resources and applications.

The current GIS organization provides service to all County departments. Individual departments also have internal resources. Some departments have "power users" who have expertise in basic GIS software and are capable of application development and sophisticated use of the GIS. Other departments are completely dependent on the services of the GIS organization within DCIS.

GIS Components

The primary GIS software selected by the County are Arc/Info and ArcView developed by Esri. Arc/Info provides tools to manage numerous data types, including vector, alphanumeric and raster data. It has allowed the construction of a County-wide distributed multi-purpose geo-referenced database. A key component of the GIS is the ability to link geographic data with other database located in external databases, like Oracle. The GIS software is used to manage, process, map and analyze all geographic information. It is also used to maintain standard maps with the addition of updates as new development takes place and other changes occur in the county. Cadastral mapping takes place as new deeds and subdivisions are recorded, utilizing Arc/Info and COGO. ArcStorm is used to house the cadastral and planimetric base maps while Librarian stores other coverages.

The digital base map consists of topologically structured digital data sets in Arc/Info native format that store points, lines, and polygons using the Pennsylvania State Plane Coordinate System, South Zone NAD 1983 and that have attribute data and/or annotation associated with each graphic feature. The legacy system of 1" = 100' mapping in urban areas and 1" = 400' mapping in rural areas has been standardized on 1" = 200' mapping county-wide in 5000’ x 5000’ grid. Plotted tax maps distributed to the 73 municipalities remain at 1"=400’ and 1"=100’.

The County obtained the following features along with the digital orthophoto data:

• DEM (breaklines and masspoints)
• Public Roads (edge of pavement)
• Driveways in excess of 200'
• Road centerlines
• Building centroids
• Major building outlines
• Hydrography
• Railroads
• Obscured areas
• Parks
• Airports

In 1995 the County began a digital tax map conversion project and hired a consultant to manage the project. The original hand-drafted linen maps were scanned into digital files and rectified to the orthophoto base. Property lines were adjusted to the planimetric right-of-ways. Buildings and natural features were used to correct questionable areas. The project was completed in 1997 and the coverages were delivered to the County. Subsequent cleanup edits have been and will continue to be on-going.

The relational database management system (RDBMS) used is Oracle, developed by Oracle Corporation. It is used to store and manage all key land record data. Oracle links all types of information stored in GIS. The primary index used to link the types of data is the Uniform Parcel Identifier, or UPI.

The tabular database has been built using a variety of sources of information. Originally some information was to be imported from the Real Estate Database (REDB) database located on the County Unisys mainframe. Subsequent project changes have identified the IAS and Recorder’s ACS as the sources of data. Information is entered into the tabular database on a daily basis as it enters the County. An example of this is when a deed is received for the sale of a parcel of land. When the deed is accepted by Recorder’s Office the tabular information that appears on the deed is keyed into the system. As the system expands numerous other user departments will enter tabular information.

Hardware

GIS and Oracle services run on a mix of UNIX and NT platforms on an Ethernet network using TCP/IP. Original equipment purchased in 1994 were IBM RS6000 250 workstations, running AIX UNIX. The production Cadastral Mapping function and GIS data are on UNIX and are maintained in ArcInfo 7.2.1 and ArcStorm. The current production Land Record Maintenance function utilizes an Oracle 8.15 database on UNIX and an Oracle Web Server front end on NT, accessed by desktop Internet browsers. Development and testing take place on a mix of UNIX and NT servers.

An IBM RS6000 J50 server running ArcInfo and Oracle was also acquired in 1994. Through upgrades over 4 years the server has 6-200Mhz processors, 2Gb RAM and 80Gb storage. The local area network solution was 10BaseT Ethernet and utilized NFS and Xwindows based client/server model. As a result of recent upgrades the network has segments of its backbone running Gigabit Ethernet over fiber, and Fast (100Mb/s) Ethernet over CAT5 UTP everywhere else. Cisco Routers have replaced 3Com. An IBM RS6000 Model 250 is the server located at the Government Services Center (GSC) location, redundantly storing GIS data in a distributed environment due to remote network limitations.

Recently 2 new IBM RS6000 H70 servers have been acquired with 2-340Mhz processors, 2Gb RAM and 88Gb storage each. One serves as a production Oracle database server (dedicating the older J50 to ArcInfo/GIS data) and the second serves as a development/test server. Six Compaq NT servers are also used as Oracle Web application, development, and ArcView distributed data servers.

The basic strategy of the Master Plan was to build a land record infrastructure centered on the parcel and its unique identifier. The first application development and maintenance operations focused on maintaining parcel data that is being shared by many agencies. The LRS functions as a source for tracking parcel history, including chain-of-title and parent-child relationships.

The 1992 Master Plan suggested 3 phases, involving 45 tasks spread out over 4 years. The first phase would be the construction of a tabular parcel database using an RDBMS. Hardware and software acquisition and a parcel data load were the next tasks. The second phase would occur when parcel data maintenance operations were in process. This phase would include the acquisition of GIS components and creation of the Registry (BLR). The next task would be to link the RDBMS and the GIS. The third phase proposed additional application development involving both the GIS and RDBMS.

The tasks proposed were implemented in a different order during implementation, due to several reasons. Essentially hardware, software, planimetric landbase and tax parcel coverages were purchased first. During this time the design and initial development of the LRS took place. In 1997 data loading was started along with testing of the system. The BLR was created and an UPI Pilot project got underway in 1998. The testing and pilot project revealed many weaknesses in the design, performance and data of the system. In 1999 a design review was performed, the results of which proposed the termination of the pilot project and a redesign of the system.

Throughout the later half of 1999 and the first half of 2000 the system underwent many substantial modifications. It is currently in user acceptance testing phase, and a pilot phase is scheduled in July and a go-live date in September.

The successful implementation of GIS requires good project management principles to successfully meet all the challenges during this process. A project of this size requires a well thought out implementation strategies with sufficient control to monitor any changes and to address any problems encountered during the implementation stage. Following are some of the problems that were encountered during the implementation of GIS at Chester County.

• Project Scope - The multi-year, multi-component, and multi-department project presented a large scope and wide breadth of tasks. Given this complexity, it is not hard to imagine that changes, problems, slipping schedules and staff turnover would make an on-time and on-budget project completion very difficult. Schedules have been delayed, and change has not always been effectively handled.

• Project Leadership - A successful GIS implementation requires a "Project Leader", who has the necessary management and organizations skills to get the job done, and to motivate people and get them excited about their work. It takes a focused leader and cooperation to overcome traditional organizational inertia and hurdles.

• Project Resources - Change occurs over time, this is a given. Ideally, conditions favoring staff longevity and retention should exist, and enable a project to be managed and executed by one group of people. When resources (staff, material, and money) are cut back, the chances of meeting the scheduled delivery are reduced. Long-term, complex projects require a steady, sustained commitment of resources. When a project’s leaders or key staff leave the project, procedures and guidelines must be in place and enforced to ensure the timely and comprehensive transfer of the project’s overall plans, intimate details and intangible knowledge from one leader to a successor. Failing this, new personnel must gain an understanding of the project from scratch. This is not always efficient or effective. Key technical staff left on a more or less continuous rotation from 1995 to 1998, until all project leaders left in the spring of 1998. New management was hired in the fall of 1998.

• Success Visibility – For a GIS project to get management and users support, it should produce deliverables and products over short time frames to allow quick results that can be seen and used. The length of the project has prevented the County, the staff and management from seeing the intended benefits of the project in a timely manner. Frequent visibility of success stories to upper management and to users is critical to keeping momentum and funding for new initiatives.

• Project Critical Path Analysis - The GIS Project would have benefited from critical path analysis during the planning and building stages. When key components and deliverables did not come on line on time, it was fruitless to continue with attempting to force the system to "go live." When schedules are not met, and major changes occur, the appropriate measures must be taken to meet the established success criteria. Change the end date, add staff, or reduce the scope, but don’t believe that the original target date can be forced when major delays or changes occur without shortcuts being taken and quality suffering.

• Quality Control - It is imperative that quality control (QC) measures must be in place and enforced through all project stages and for all key deliverables. Clearly identify the specifics of the delivered product, including QC criteria to be met prior to purchasing data. There are several problems with delivered GIS data now and the required post-delivery QC efforts are much more costly and lengthy.

• CASE Tools – Use computer-aided software engineering tools for the design, analysis and building of applications and databases. Use business process and data model diagramming tools. Many hours of development will be saved by use of a tool such as Oracle Designer.

• Outsourcing Issues – Carefully review needs assessment reports and master plans.  Make sure the content of each is addressing your needs.  Establish criteria before hand on what your organization wants from an assessment report.  Often a consultant will reuse documents prepared for other clients, with 60% of its content being off-the-shelf, boilerplate, this-is-what-GIS-does material. Hiring consultants requires a full time staff to manage and monitor. Consultants generally don’t take ownership of a system or pride in searching out issues and imperfections that aren’t in their scope. Consultants generally sell you the idea of their expertise, and then insert less experienced personnel to learn on your project, at your expense. Make sure the GIS consulting firm you hire has real world IT and database experience, and has a successful track record in implementing systems. Forty percent (40%) of all IT and 60% of GIS projects never reach fruition because of cost, schedule overruns or poor performance upon delivery.

• Prototype software – Make sure the final system delivered isn’t the prototype with some window dressing added. At some point during development the prototype must be scrapped and the final version built with all the lessons learned. If the comment "We’ll fix this in the next release" is made too often, it might be time to scrap the current version or the developers responsible.

• Business Procedures Rules Definition and Data Model - Prior to acquiring and implementing a new technology solution, ensure that adequate business process reengineering has been successfully completed. Proper business and data modeling, procedures and business rules must be documented and signed off on by the parties involved. Otherwise, the benefits of the new technology will be wasted as people go about doing things the way they have always done.

• Project Management – Having a good project management process in place helps avoid excessive workloads, cost overruns, changes in the project scope, work loss, redone work, insufficient resources and missed deadlines. Close the loop on controlling the project execution, manage changes, and follow through on quality control. Without it, all of the aforementioned can and will occur.

Enterprise-wide GIS requires a number of things to be planned and obtained. A sound technology architecture with a well-designed environment, adequate network connectivity and reliability, documented maintenance, backup and recovery procedures, and an appropriate number of trained staff is only one component. Well-defined business functions, procedures, rules and data models are critical in the design of applications and how they perform during daily operations. Communicating with and understanding end user needs and expectations is important to the success of the plan. Whether designing a system or an implementation plan, it is important to consider the following guidelines:

Paul R. Dioguardo                                                   Mohammad Tariq

GIS Technical Services Manager                          GIS Project Leader

pdioguardo@chesco.org                                       mtariq@chesco.org

Chester County. "Chester County Land Records System Master Plan Update."

PlanGraphics, Inc.: 1992.

Chester County. "CHESCO-LRS Re-engineered Land Records Logical Design."

Chester County DCIS/PlanGraphics, Inc.: 1996.

Havan-Orumieh, Marina, et al. "Overview of the Reengineering of Land Records Process in Chester County, Pennsylvania." 1997.

Chester County. "Land Record System, Project Review and Recommendations"

Dioguardo, Paul R., Brinton, Patricia, Chester County GIS, DCIS staff: 1999