Data Modelling is the means by which we develop and communicate an understanding of the information needed to describe a topic in order to support a business activity. The data model describes the things of interest, the relationships between them and the data quality rules that govern the acceptable values and relationships.
Data may be modelled using traditional entity relationship notations or may be extended using UML notations to include object behaviours as well as their data properties; an approach that is particularly popular with more Object
Oriented systems such as ESRI’s ArcGIS Geodatabase and GE Smallworld’s version managed datastore.
Although the modelling of data for conventional Relational Database Management Systems is well understood, the modelling of the spatial aspects of information in both the operational and warehouse contexts varies considerably among practitioners. Modellers need to understand the nature of spatial data at a variety of scales as well as spatial relationships and associated modelling techniques to support linear referencing and topologies.
In order to promote a Logical data Model to a Physical Data Model for a specific platform, modellers must possess an expert understanding of the spatial implementation of that platform in order to support rules management, long transactions/versioning, historical data management, disconnected/mobile editing and performance tuning. The performance tuning of spatial databases is essential to success of the implementation and requires a detailed understanding of how the platform indexes spatial data, the data access patterns of the key applications, and may also require knowledge of application network use characteristics, database block and page sizes and partitioning options.
Data never remains in one database for long and all solutions should be designed with integration mind from the beginning. Data updates may originate from multiple sources, and conversely it is highly likely that other systems will have an interest receiving data updates from the new model. Modellers should be familiar with data exchange patterns and protocols, particularly XML/ GML web services, data replication and messaging infrastructures.
We have substantial experience in modelling spatial data for a number of domain areas including water, electric and cable utilities, land management, environmental management, asset management, network planning, transportation, etc. We have expertise in numerous geographical information systems and spatial data management technologies, as well as the modelling tools (Oracle Designer, Rational Rose, Visio, ERwin, Power Designer, Objecteering) and techniques relevant to each.
Spatial data profiling is often performed in conjunction with a modelling effort. Profiling involves conducting hands on investigation of the actual data contents in order to discover patterns within data sets to identify potential relationships and data integrity rules. Spatial Consultants can examine existing data holdings to reverse engineer data models or to assess compliance of the data with quality specifications. We are familiar with numerous toolsets such as SQL, FME, 1Spatial Radius Studio and numerous data warehousing and transformation tools