Zero to Enterprise in 24^th Months

Lessons Learned From Building Enterprise GIS at the Port of Portland

Presented by Preston Beck

Abstract
The Port of Portland owns and operates marine and aviation facilities and a significant amount of industrial property.  It is a large, mature, diversified, and geographically dispersed organization.  A little less than two years ago, only a handful of people at the Port knew what GIS was.  Today, the Port has an Enterprise GIS that is accessible to everyone who works there.  This system is used for a wide variety of applications ranging from planning and infrastructure maintenance management, to marketing and environmental management.  This presentation will provide a managerial perspective on how this system was planned, designed, developed, and deployed, including technology selection and implementation, data conversion/acquisition/creation/maintenance, organizational design and staffing, training, and procurement

Introduction
Zero to 24 months may not be too fast when you’re talking about sports cars, but for an organization implementing an enterprise new GIS it’s pretty speedy.  Two years ago, the Port had no GIS capabilities within the organization, save for a few isolated GIS renegades. Then in a fairly quick project, rolled out an enterprise GIS that is available to every employee.

The Port of Portland is Oregon’s largest Port authority. It owns and operates one international airport, three general aviation airports, five marine terminals, and a significant amount of industrial land. The Port’s holdings are geographically dispersed throughout the Portland metropolitan area.  The Port’s GIS will be used for a variety of applications ranging from land-use planning and infrastructure maintenance management, to marketing and environmental management.  This paper provides a managerial perspective on how this system was planned, designed, developed, and deployed, including technology selection and implementation, data conversion/acquisition/creation/maintenance, and organizational design.

Project Evolution and Impetus
The idea of implementing GIS at the Port was not new when the project officially started in 2000. Early attempts were started nearly a decade ago, with the Engineering Department taking an interest in GIS as an additional tool to augment their established CAD environment.  The Port participated in multi-agency efforts in the early 90’s when GIS was beginning to take a stronghold as a data management tool in governments, but failed to ever make any substantial commitment to GIS. This was due to a couple of reasons. No united effort to implement GIS across the organization existed.  The only real interest was from one department, Engineering, and the operating departments were not included (i.e., Aviation and Marine Departments.). Also no cost benefit analysis for GIS was ever performed to show how it would eventually provide a return to the Port. Finally, and most importantly no top level management support within the organization existed, making it difficult to secure any funding for development and implementation.

While other agencies in the Portland area moved forward and invested in GIS, the Port remained on the sidelines, continuing to use its CAD as its primary repository for geographic information. But while the Port may have made the corporate level decision to not invest in GIS, interest by individual departments continued to grow. Eventually during the mid 90’s some departments became GIS renegades, purchasing their own ArcView licenses and acquiring data layers from various sources.

With the interest growing and GIS data issues multiplying, the IT Department decided to initiate a feasibility study of GIS that involved a cost-benefit analysis. A strong concern of IT was that if the organization as a whole didn’t do something about GIS, departments would most like continue to develop their own systems independently, a maneuver that could have serious negative repercussions, from an IT perspective.  At the same time, by coincidence, several top level managers began to support GIS and the feasibility study.

At the completion of the feasibility study, several elements came together which ultimately lead to the decision to make GIS a Port-wide, or enterprise initiative. These include:  1) a cost-benefit analysis that demonstrated a clear return on investment, 2) support from the major participating departments, 3) funding, and 4) a project sponsor. Without these four elements, it is likely GIS might have continued remain a low priority for the organization for the foreseeable future.

Feasibility Study Outcomes
The feasibility study contained elements that assisted management in making the decision to fund an enterprise GIS implementation.  The study first identified fundamental problems with the Port’s management of spatial information. These were:

·        No consistent process or system for managing and maintaining mapped information

·        No system in place to provide a computerized database for capture, storage, retrieval, analysis, and display of geographic data

·        The existence of redundancy and/or gaps in obtaining, maintaining, and sharing geographic data sets

·        Inadequate means of geographic data sharing with other agencies and contractors

The study also noted several opportunities for the Port related to the management of spatial information. These included:

·        A strong interest in using a map interface to geographic information

·        Strong regional support and extensive use of GIS technology

·        Cost savings to implementing a Port-wide GIS

The cost savings for the Port was of course the most important determiner of implementing GIS.  Cost-savings would result from:  better data management, a reduction in data re-creation, a reduction in data duplication and verification, reduced data purchasing redundancies, and a reduction in cumbersome data retrieval. After the initial investment in the core infrastructure and 5-year payback, GIS was would provide approximately a $100,000 annual savings. And it is likely the benefits would increase as further GIS development occurred.

In addition to the strict cost benefits to GIS, there were strategic benefits for the organization.  GIS is empowering technology, allowing for better access to information, therefore it would allow the organization to improve its regional leadership for transportation-related issues, manage environmental impacts of Port operations, be better and more informed decision-makers on issues affecting the Port, and provide better customer service.

Project Formulation
With the GIS capital project officially established and a project manager devoted solely to the implementation of GIS, the first step involved forming a project team and establishing project goals and objectives. Project goals and objectives were essential to keeping the project stakeholders focused and preventing the constant battle of scope creep. The Port GIS Project had one single goal and a set of overall project objectives as well as technical objectives.

Project Goal:
Establish core GIS infrastructure for the Port of Portland

Project Objectives:

·        User needs/interests should drive project

·        GIS must be made widely available

·        GIS needs to be easy to use

·        Initial efforts need to focus on core GIS infrastructure

Technical Objectives:

·        GIS must be an enterprise system

·        GIS needs to be consist with IT Strategic Plan

·        The GIS database design must allow for efficient data sharing

·        Maintenance and updates need to be as seamless as possible

·        Design Port GIS to be flexible to allow for future customization

Lesson # 1 Establish clear realistic and resource backed project goals and objectives. Stick to them!

Project Team Structure
The Port of Portland follows a Project Delivery System model for projects. This framework relies on a project sponsor, the owner of the project, and the chartering of a team of stakeholders.  In the case of the GIS Project this involved two project co-sponsors, both directors, a steering committee, composed of representatives from around the Port, and a technical team, composed primarily of IT Department staff. The GIS steering committee members were carefully chosen for their ability to act as steadfast decision-makers and be good communicators, both up and down the management structure. In other words, project champions.

Steering Committee members played a vital role in keeping the project visible. The steering committed proved to be extremely important given that the knowledge of GIS was so limited in the organization and the project manager’s time limited in being able to disseminate project information.

Lesson # 2 Establish a steering committee representing all stakeholders, seed with reliable enthusiastic project champions.

Project Structure
The project structure utilized standard IT or IS implementation planning methodology for the development and implementation of enterprise GIS. The basic outline is show in the accompanying chart.

Although the Port had gone through a feasibility study, where a high level needs assessment had been completed, it was still necessary to revisit the needs assessment because of a couple reasons. The original needs assessment had a department “silo” approach and failed to address the business processes of the Port. For example, the original feasibility study contained information about Marine Department needs and Aviation Department needs. While these are two very different departments within the Port, they both have similar business functions, such as operations and maintenance of their facilities. Reconfiguring the approach to be in line with the business functions proved to be more efficient in terms of looking at similar functions of the departments. Additionally, these business functions crossed to other departments and therefore they needed to be included in the process. For example, the Environmental Affairs Department is part of operations and maintenance of Port facilities. During the needs assessment workshops it was critical for all these participants to be at the table when discussing the needs of the specific business functions.

Lesson # 3

Based needs assessment on business functions rather than traditional department silos.

User Topology -Client Framework
Based on the Needs Assessment, it was clear that several different categories of people would interact with GIS data at the Port.  These different types of users would need different types of software applications and training.  The types of clients groups are:

Doers:
System administration, high-end maintenance/editing

Users:
Advanced display, query and analyze spatial databases. Advanced analysis/modeling, cartography, and quality map production. Create,   manipulate, and display spatial data in preparation for analysis and mapping.
Viewers:
The largest set of GIS consumers at the Port.  Viewers primarily browse and view “ready-to-use” spatial data for their particular needs, mapping and limited analysis via pre-determined parameters.  

The client framework is set up like a pyramid with the Doers representing the smallest group of users and the Viewers at the base of the pyramid representing the largest group.  This framework model was very important during communication efforts to the departments. It allowed staff to see not only the type of software they would use, but also understand the overall structure of the Port’s emerging GIS.

Lesson # 4 Establish a topology of users. Create a client framework and let this be the primary driver of your software mix –User driven!

Technology Selection
During the feasibility study, the Port went through a vendor selection process. GIS products from Environmental Systems Research Institute (Esri), Intergraph/Bentley, and MapInfo Corp were compared during the study. It was determined that in terms of range of products, functionality, and matching up with Port-wide needs, including web-ready or web-enabled software for GIS, Esri products would meet the range and complexity of GIS requirements at the Port.  In addition, it was noted that additional third-party software might be practical for customizing commercial software (COTS).  The “pre-selection” of the vendor helped speed up the infrastructure project, allowing the team to focus on determining the proper suite of Esri products and appropriate third-party software.

Database Design
The database design ended up being a relatively uncomplicated process. This was due to a couple of reasons. A large majority of the data used in the Port’s GIS is from external sources, therefore design work was negligible since the Port merely purchased into a subscription and pre-established data set.  Another reason was that several of the systems that had been originally desired for integration into Port GIS, such as the work order systems and lease managements systems were omitted because of internal disputes or extreme costs of integration.  For the remaining internal data, the Port used the Spatial Data Standard (SDS), established by the CAD/GIS Technology Center. 

An interesting challenge for the Port GIS database design is the variety of data. Port staff desire data from a variety geographic areas. For example Aviation and Marine marketing staff need access to global data sets, planners need access to data for the Portland metropolitan area, and operations staff desire access to facility information.  Since the data from the different geographic areas has varying projections, and accuracies, the logical arrangement was to develop “geographic nests”, or a 4 tier data structure, World, Northwest, Metro, and Port.  The different categories also allow the Port to easily integrate external source data structures.

Lesson # 5 Use established database design frameworks whenever possible.

Data Conversion, Acquisition, and Maintenance
Data conversion often is the most difficult part of any GIS implementation project.  Fortunately, a rock solid needs assessment process helped keep the project team focused on prioritizing potential conversion data. The Needs Assessment step established Tier 1 –high priority and Tier 2 –nice to have data.  An early decision by the team was to put Tier 2 data immediately on the back burner, as the list was long and much of the data was either on hard copy maps or not even available (data collection was not in the scope of work).  In order to stay within the budget, it was necessary to establish guidelines for prioritizing the remaining Tier 1 data.  Below are three levels used to prioritize the various conditions of spatial data.

Level 1

·        Data is in GIS format (shapefile or ArcInfo file format)

·        Spatial reference meets standard

·        Metadata available

Level 2

·        Spatial data is in electronic format (such as CAD, DGN)

·        Spatial reference is different than standard

·        Attributes are in some database and contain unique ID

·        Tabular data is relatively clean

·        Owner identified for maintenance

Level 3

·        Spatial data is in hard copy

·        Attributes have no unique IDs or are not in database

Level 1 data would be included in Port GIS.  Level 2 data needed some level of conversion and preparation work and would need to be further prioritized Level 3 data would not be included.  

Level 2 data ended up being primarily engineering utility data which is maintained in Auto-CAD.  The majority of the conversion work focused on these utilities.  Here again, prioritization was important given the limited amount of resources.  The Steering Committee along with key Engineering staff helped prioritize the utility information.

Insisting that an owner be identified for each layer became a crucial guideline.  Before conversion was undertaken, the owner had to identify who would maintain the data and how often they intended to update it.  This proved to be a good mechanism for keeping data conversion in control.  Too often during the project the appetite for more GIS layers overshadowed the responsibility of maintaining the data in the future.

Lesson # 6 Establish guidelines for prioritizing your conversion lists.  Realize that that they are only guidelines and individual calls will still need to be made.

Lesson # 7 Establish clear ownership and maintenance of data layers.

Application Design
Application design for Port GIS proved to be a relatively straightforward phase. As mentioned before, Esri had been chosen to be the vendor of the software allowing the infrastructure project to focusing on establishing the exact mix and design of the suite of products. Further guiding the decision-making was the technical objectives of the project, specifically the consistency of the IT Strategic Plan which calls for common systems and architectures installed package software, and standard software components. Creating a system with extensive customization involving programming resources was discouraged.  

The mix of software reflects the Port GIS client framework previously discussed. At the “Doer” client, ArcView 8 is used for the administration and high-end maintenance. The new editing tools provided in ArcView 8x are a significant improvement over ArcView 3.x, enough that purchasing ArcEditor or ArcInfo was not necessary. Additionally, the Port has no legacy ties to coverages so there was no need for coverage editing. Furthermore, most of the Port’s internal drawing work is done in CAD and converted to GIS so the existing CAD software is quite satisfactory.

For the “User” client, the obvious choice was ArcView. The Port is using 3.2 at this time. The jump right into ArcGIS for an organization brand new to GIS would have been overwhelming, not to mention a strain on resources, so this was thought to be a safer strategy. In the next year or after the Port learns to walk with GIS, it can “run” with ArcGIS.  Most certainly the Port will want to do this before too much customization occurs using Avenue as this is not carried through in the ArcGIS world.

“Viewer” clients need a GIS that’s easy to use and requires minimal training. With one of the project objectives to make GIS widely available, the logical choice was an Internet Map Server (IMS). The decision on this was made easier by the technology advancements that have occurred in the last several years.

While ArcIMS was the preferred choice for an internet map server, the Port wanted to be able to incorporate additional functionality from such as linking the IMS to various databases and it wanted to do so without intense hard-coding.  GeoNorth’s MapOptix, an administrative extension to ArcIMS proved to be the best third-party solution for building a robust IMS without the major programming headaches.

Organizational Design, Selling the Concept of Enterprise GIS
The organizational design of Port GIS is one of the most important critical success components of its existence.  GIS was approved as a capital project based on the assumption that it would be administered by IT and upon project completion woven into its application portfolio of corporate systems. The initial Director level approval however didn’t necessarily guarantee that knowledge of the new corporate system would trickle down and receive acceptance by department level management, therefore it was necessary to develop a communication plan to inform all levels of staff of the concept. This included e-mails, department presentations, brown-bags, and bulletins. The primary message –Port GIS is corporate system administered by IT but data ownership resides, just as it does now, with the individual departments. 

Although the concept of administration and data ownership seemed very basic, especially to staff directly involved in the project, it often failed to register in the minds of several staff, and thus it was necessary to be always proactive about walking people through the corporate concept framework.

Lesson # 8 Despite your best intentions and laid out plans, some people still will not get the concept of an enterprise GIS, be proactive about communication.

Conclusion
Building an enterprise GIS for an agency is a Herculean undertaking.  The Port met the challenge however, by providing a very usable, widely available, and “easy to get to” GIS.  Core infrastructure deployed in this project should provide Port of Portland staff with a substantial base for future growth in the GIS world.  Hopefully the lessons discussed in this paper will provide readers with value tips if they might be considering implementation of an enterprise GIS.

Acknowledgements
Special thanks to Port GIS Project team members: Roger Crystal, Carol Baumann of Titan Corporation; Marshall Payne, Andy Sloop, Nick vanVeldhuisen, Gerald Pisarsky of GeoNorth;  Dick Bolen, Mark Bosworth, Alan Holstead of Metro, and Port of Portland staff, too numerous to identify here.

Preston Beck
GIS Project Manager
Port of Portland
PO Box 3529
Portland, OR 97208
503 944-7514
beckp@portptld.com