The City of Portland in Oregon, USA, comprises a number of divisions or bureaus, each responsible for the administration of a specific area of local government, e.g. buildings, police, water, planning, transportation, etc. Over time each bureau had adopted and developed a set of IT/GIS solutions to meet their needs – this worked reasonably well in isolation but over time led to a number of significant problems:
- Organisations found it difficult to share information of common interest;
- Multiple organisations were maintaining private versions of key datasets e.g. road network, land parcels, and addresses;
- Users found it difficult and time consuming to locate (and obtain) the most accurate or upto date version of any given dataset;
- Decision makers could not access the ‘big picture’ as datasets could not be integrated easily; they therefore lacked confidence in the data that supported them;
- The city was investing in redundant application development as similar bespoke functionality was created for each platform.
The solution was to enable citywide data sharing through a GIS Hub system that enables bureaus to publish their data into a corporate data structure in a common format, and to extract data provided by other bureaus in the working format of their choice.
Business Architecture
The City centralised spatial data management into a corporate body (CGIS) which was responsible for coordinating investment and development in GIS data and toolsets. A business case examining the expenditure on redundant application development and data maintenance as well as time wasted by users in locating and verifying data clearly justified this approach. CGIS contracted Convergent Group to support them in developing a GIS infrastructure to enable spatial data sharing; Graham Morgan was engaged as the Project Engineer for the design and implementation. The ability to share data required that a number of data governance issues be reviewed including data ownership and maintenance responsibility, data standards for enterprise-wide use and corporate data funding.
Data Architecture
The GIS Hub was attractive to city bureaus as it enabled them to continue to use their systems of choice to manage their data – but to also benefit from datasets managed by other bureaus or external organisations. The data could be consumed in the format and projection of choice but easily shared by the GIS Hub infrastructure. A central data model was developed for the Hub into which the source data could be translated and transformed. The central data model also provided a platform for enterprise reporting as all data was now available in one place in a common integrated format and coordinate reference system.
Technical Architecture
The GIS Hub was developed using ArcSDE with the data stored in Microsoft SQL Server. The data was translated and transformed between bureau systems and the Hub by customised FME. Data was bulk loaded into a staging database where data integrity checks were performed before the data was published to the Hub. A custom FME application was used to translate from a range of formats into and out of the Hub; this facility was made available on the corporate intranet so that data requests could be queued and processed as received. The Hub was implemented on three SQL Server databases and the spatial data was automatically replicated between them to support local performance and general availability. A number of technical challenges were overcome during this project which involved working with ESRI to develop the (then) ArcSDE data structure to support database replication and declarative SQL. The CGIS team developed a cutting edge GIS web site based on Cold Fusion and GeoNorth’s MapOptix (and later ArcGIS Server), that accessed Hub data and made it available internally and externally. The website may be accessed online at: http://www.portlandmaps.com/