This paper is a case study in action. The City of Pasadena is located in South Eastern Harris County, a ‘suburb’ of Houston. However, Pasadena is not a typical homeowner suburb (it is highly industrial and has a small mobile population), and like many cities of similar size (150,000 population) and with a proximity to large a metropolitan city; we are expected to deliver ‘big products’ on a ‘little budget.’ Thus is born the notion of re-engineering. The paper will elaborate how a mid-size city can successfully implement ArcGIS 8 from the ground up. It will further document how to address the needs of the users while maintaining an ‘infant’ GIS, and provide hard results to the city leadership (Mayor, Manager, and/or City Council). Re-Engineering will document the rites of passage from conceptual design, to department use-case analysis, to implementation and pilot projecting to the actual step of re-design and re-implementation. It will prove that this process can be efficient and successful, and maintain that mid-size cities can be self-sufficient (i.e., without the aid of consultants).
Project and scope creep is, has, and always will be a part of bringing up any system. However, when a project goes through a total metamorphosis mid-implementation, a new term is applied. Born is re-engineering or the changing of business practices both standard and custom, to fit a new theory or technology. Re-Engineering as applied in this scenario will document the rites of passage in implementing an ArcGIS 8 system from conceptual design, to department use-case analysis, to implementation and a pilot project, to the actual step of re-design and re-implementation. It will prove that this process can be efficient and successful, and avoid many of the pitfalls that can cripple an otherwise nonflexible project.
For the case study, this paper will follow the City of Pasadena’s implementation of ArcGIS from GIS infancy to a fully installed and functioning geographic information system composed of ArcGIS 8.2 in various flavors and ArcSDE running on Microsoft SQL 2000.
The City of Pasadena (herein the city) is located in south eastern Harris County, suburb of Houston. However, Pasadena is not a typical homeowner suburb (it is highly industrial and with a high percentage of surrounding communities commuting to it), and like many cities of similar size (150,000 population) and with a proximity to large metropolitan city; the city is expected to deliver ‘big products’ on a ‘little budget.’
The need for a more data rich mapping program, and with GASB 34 looming around the corner, the time was right to implement a robust GIS program. A composite team was organized from the Computer Information System department and the Public Works’ Engineering department to facilitate discovering the need.
GIS is just an acronym without understanding what it stands for. In April of 2001, there were wide eyes and hopeful aspirations for a new geographic information system, and yet with great angst the city met with our Esri representative, David Dignum, to discuss the future of the project. With any project or system, the initial meetings to discuss the project are critical, and these meetings were no different. Esri was quick to explain the nature of the partnership they would have with the city. Note here, that this relationship was a partnership in every sense of the word. David and his colleagues at Esri understood how to make a GIS work, we as a city and as a team needed to quickly learn how to implement it and how to make it work.
Out of the initial meetings our GIS team was created. Staff members from engineering, planning, and information services came together to formulate what exactly we planned on getting out of GIS. At this point two options are generally available to medium sized cities; either A. employ a consultant or B. do from within (noting it is most likely that there is not a GIS department already within a city). Both options are valid and arguments for both can be made, but regardless a GIS team at the City level must be created.
The City of Pasadena choose to do from within; with that and with a dedicated GIS team in place we were prepared to take the next step.
Imitation is the sincerest form of flattery, and often the best way to implement a system. In this case, we were a long way from imitation, we had not settled on the direction to carry our project. Specifically, we had not decided which to implement ArcGIS 8.0 or ArcView 3.x. Both products had advantages and disadvantages, and our choice was as much internal politics as it was system functionality. Such decisions should always be expected, and throughout the implementation of the system, be prepared to deviate from what seems at the time to be the best solution.
In the Houston – Galveston user community at that time, there were few to no full installs of ArcGIS 8.0, and for that matter not a lot across the country. We searched the internet, we contacted the Conference of Governments for our area, the Houston Galveston Area Council, we contacted other municipalities, and attended every regional meeting we could find. At each of these meetings we observed and realized that the decision to implement an ArcView 3.x system would be far easier because of the built in support, but if we wanted to take advantage of newest technology we would have to pursue ArcGIS 8.0.
After lengthy meetings and long discussions with the vendor and other involved parties the decision was made to install ArcGIS 8.0. We planned to capitalize on the newest technology and to install an ArcSDE server and run our entire GIS data from geodatabases. Sounds good doesn’t it? It did to use too.
After making our choice of software platforms, the city had a few other wrinkles to iron out, namely how do we implement and maintain a completely new GIS system. Again we turned to the user community and the vendor. The GIS team dedicated our initial meetings to the needs of engineering, planning, and information services. The following criteria were established;
Engineering:
Data
• Street center line map to serve as basemap
• Utility maps to be converted from Microstation drawings
Hardware
• Outdated drafting stations converted to GIS grade workstations
Other
• Prepare the system for other applications
• Train drafting techs to use ArcINFO
Planning:
Data
• Obtain parcel map and necessary Census info
• Change internal policies regarding external data submission
Hardware
• Acquire GIS grade machines
Other
• Prepare for upcoming redistricting process
• Train in ArcGIS software, regardless of flavor
Information Services:
Hardware
• Spec and purchase necessary workstations for all departments involved
• Spec and purchase suitable server for ArcSDE
Software
• Purchase all licenses for ArcGIS software
• Install and maintain software
Project Management
• Serve as project manager for implementation of ArcGIS
• Arrange all training
• Propose pilot project
Without employing a consulting firm to draft the implementation plan, the bulk of the work is finding needs and proposing solutions to problems. Logically, in the city’s estimation, the best place to start was with cities of similar scope and complexity to Pasadena, examine their GIS programs and implementation plans, and develop our own from there. Most cities were forthcoming with information and in general the user community embraces any request for information we would put out, however, the largest road block we encountered was the lack of previous ArcGIS 8.0 installations.
The development of the implementation served a threefold purpose:
1. It serves as a guide for the GIS; ideally, it is updated and referenced when ever a major system change or sub-system is installed.
2. It proved to be a valuable training tool internally. The first part of the document outlined basics of GIS and the vocabulary. Basic understanding of terms and creating a working knowledge of the hierarchy of products can often be a monumental undertaking.
3. The implementation plan and its execution gave legitimacy to the project throughout the pilot project phase as well as changes in city leadership.
The conversion pilot was ruled out early, as the city had secured a street center line map with address points attached from the Geographic Data Committee of Houston Galveston Area Council (of Governments).
Fundamentally, the neighborhood pilot had less risk and at the same time less reward. While it would have shown success, it may or may not have shown the full functionality of the software as well as its ability to save time and money.
During the deliberations over which pilot to undertake, the Director of Planning approached the GIS team and stated that the city had to begin the process of redistricting the city council districts, a process that was typically long and very arduous. In the past, the Mayor would appoint various people to a standing committee to redistrict and would also employ outside consultants to do much of the labor. In turn, the planning department would do much of final computations on top of the daily activities. This was a perfect opportunity of a fledgling GIS program. We would be able to save the city an enormous amount of money as well as expedite the redistricting plan.
The pilot was chosen, and the GIS team sub-divided and introduced new players from several new departments. Now representatives of the Mayor’s office as well as other departments would have input into the pilot project. Thus, by the selection of this pilot the project grew.
City wide initiatives are rarely simple and by adding in something as volatile as politics, the degree of difficulty can be increased greatly. We established a few guidelines to make the project go as smoothly as possible. First, let the analytical data speak for itself, do not (at least) initially become subjective with the placement of districts. Second, put as little human interpretation into the data as possible. Take the data apply it to a map, and then create the districts. Do not create districts and try to make the numbers match the lines. Third, the Mayor had final say in anything. This rule might seem to be a given, but we wanted to make sure we knew who was running the show.
By City Ordinance, once the Census data is obtained from the Census Bureau, the Mayor has 6 months to propose final redistricting map to City Council for approval. Within this time table several steps have to be made:
1. Cleanup the data obtained from the Census Bureau
2. Transfer the data into workable tables
3. Apply the data to the basemap
4. Create a working redistricting map
5. Allow for modifications and meetings to facilitate those modifications
6. Mayor’s approval or re-modification
7. Individual Council review
8. Submission to Council for vote
In the past, a committee was established within days of receipt of the data from the Census Bureau, and a working plan would normally take 4-5 month to create, then any modifications would be last minute. All in all, the process would take 14-16 people 6 months to complete.
The entire process from cleanup to submission to the Mayor for his approval took 3 people 4 days to complete. This quick turnaround gave the Mayor the freedom to work the map in ways previous mayors had never been able. In council review, most Councilpersons had no problems with the map, noting that it was the fairest map created in many years. Those having problems quickly learned that the analytical data had been applied analytically, and there was little room for political interpretation.
The results of the redistricting process can be found here(http://www.ci.pasadena.tx.us/redistrict.htm).
The success of the pilot project created opportunities for advancement of the project as a whole. The project was expanded by the hiring of another GIS technician as well as employing a paid part-time intern to help with many of the drafting responsibilities. The city expanded is licenses to 4 ArcINFO, 2 ArcEDITOR, 5 ArcVIEW, and ArcSDE on SQL 2000.
The sudden growth also revealed a major problem in our implementation plan. The implementation plan had been constructed with the data sources available at the time, namely previous plans that had implemented 3.x. While, many of the project management theories carried over, many of the data structures did not. The city’s GIS division had executed the pilot project using layers, shapefiles, INFO tables, and exchange files. This was an accurate way to attack the pilot project in ArcView 3.x, however we needed to convert our data and start looking at geodatabases if we were to take full advantage of the ArcGIS 8 power.
Saving this discussion of another time, it is sufficient to say that the data structures between ArcView 3 and ArcGIS 8 are significantly different. While ArcView 3 data structures can be used in ArcGIS 8 applications, the true power lies within converting data into geodatabases and letting the data reside on a SDE server. The city understood this, and following its initial reasons for going with ArcGIS 8, proceeded with the data conversations and retraining and ultimately, Re-Engineering.
It is critical to say that re-engineering is not a sign of failure, nor is it a sign of unrecoverable mission or scope creep. It is actually a sign of success; it shows that the project is able to cope with significant change and flourish rather than fail. Many projects die at this point, and many dollars and man-hours are lost. Therefore, it is important to establish within the implementation plan, that re-engineering is acceptable and can be expected.
For the city, re-engineering was a by-product of moving ahead so quickly. After the success of the pilot project, the GIS project was funded and it allowed for further development and training. Therefore, we slowed development down and focused on training.
Fundamentally, the creation of shapefiles and datasets is analogous between ArcView 3 and ArcGIS. However, the added dimension of the geodatabase and geometric networks creates a significant learning curve.
The city was able to train technicians in Arc fundamentals, but even after 40 hours of geodatabase, the basic concepts were all that we were able to grasp.
It took the true implementation of geodatabases to understand their structure and ultimately there power. The GIS team was able to organize large amounts of data, placing it into carefully arranged databases, and we allowed for the modification of that data across the city. This was an unrecognized benefit of re-engineering; because of the expertise we had in GIS already, the transition to geodatabases was quick and our ability to manage the data increase significantly.
As with any system, the data is only as good as what you do with it. A GIS is no different. However, the overhead work is significantly larger with a complex GIS, the creation of data, the acquisition of specialized training, and the day to day upkeep of data is extremely difficult process, and one that can change at any time. Once the foundation of the GIS is laid, the dividends are ready to come in, and one of the best investments you can make in your data is a secondary system to enhance the data.
In our city we are currently implementing two such projects, a public works work-order maintenance system and a public safety management system (with GIS based crime analysis). Each system with capitalize the data within the GIS, each department will be able to utilize data never available to them before the implementation of a GIS. Ultimately, work will be achieved more efficiently, problems will be solved faster, and data will flow freer because of a geographic information system.
Currently the city has implemented the following geodatabases completed:
• Streets (geometric network)
o various intersections defined as objects
• Storm sewer
• Water Distribution
o joints defined as objects
• Traffic network
o traffic control devices
o traffic counting devices
• Address points
• Arial Photography
• Various others in near complete form
In the future, we are looking to expand our capabilities into ArcIMS and take advantage of future regional data sharing capabilities. Until then, we will continue to advance our GIS project and re-engineer when necessary.