Spatio-temporal evaluation matrices for geospatial data

Abstract

The global geospatial community is investing substantial effort in providing tools for geospatial data-quality information analysis and systematizing the criteria for geospatial data quality. The importance of these activities is increasing, especially in the last decade, which has witnessed an enormous expansion of geospatial data use in general and especially among mass users. Although geospatial data producers are striving to define and present data-quality standards to users and users increasingly need to assess the fitness for use of the data, the success of these activities is still far from what is expected or required. As a consequence, neglect or misunderstanding of data quality among users results in misuse or risks. This paper presents an aid in spatio-temporal quality evaluation through the use of spatio-temporal evaluation matrices (STEM) and the index of spatio-temporal anticipations (INSTANT) matrices. With the help of these two simple tools, geospatial data producers can systematically categorize and visualize the granularity of their spatio-temporal data, and users can present their requirements in the same way using business intelligence principles and a Web 2.0 approach. The basic principles and some examples are presented in the paper, and potential further applied research activities are briefly described.

Introduction

The quality of geospatial data is a measure of the difference between the data and the real world that they represent. The greater this difference, the poorer the quality of data and the smaller its true value. It is not possible to create a perfect representation of the infinite complexity of the real world (Goodchild, 2006). Spatial data is inherently uncertain (Couclelis, 2003) and it is inherently likely that the same reality is modeled differently by different people (Bédard, 1987).

In describing data quality, there are many dimensions of data at different levels of detail and also many different components, known as quality elements, in the standards. Setting and implementing international quality standards is most important in various aspects, especially in improving quality, safety, security, environmental protection, and economical use of natural resources. The standardized descriptions of the eight basic quality management principles as they appear in the ISO 9000 (ISO, 2005a) and ISO 9004 (ISO, 2000) standards are:

• Principle 1: Customer focus.

• Principle 2: Leadership.

• Principle 3: Involvement of people.

• Principle 4: Process approach.

• Principle 5: System approach to management.

• Principle 6: Continual improvement.

• Principle 7: Factual approach to decision making.

• Principle 8: Mutually beneficial supplier relationships.

These principles provide conceptual guidelines for any business. Furthermore, the International Organization for Standardization (ISO) with its “Technical Committee TC 211–Geographic information/Geomatics” has developed the ISO 19100 series of domain-oriented standards in the last decade, which are specifically dedicated to geographic information and its quality (ISO, 2002, ISO, 2003a, ISO, 2003b, ISO, 2005a, ISO, 2005b, ISO, 2005c, ISO, 2005d, ISO, 2006, ISO, 2007a, ISO, 2007b, ISO, 2007c, ISO, 2008, ISO, 2009a, ISO, 2009b). These standards address the basic issue; that is, how well the geographic world is represented by the data. The basic goal of these standards is to enable interaction of geospatial datasets between different data models and different applications. The importance of user awareness of where and how specific datasets can be used in their applications is growing with the growing number of different data models and different levels of quality of geospatial datasets. Among the important aspects of these standards is the efficient and effective global dissemination of geospatial data and technologies as well as good practices, thus contributing to global economic and social progress.

Among these standards, ISO 19113 (ISO, 2005b), ISO 19114 (ISO, 2005c), and ISO 19115 (ISO, 2005d) establish principles for describing the quality of geographic data in the quality-evaluation process and summarize quality into the following well-known elements:

• Positional accuracy.

• Temporal accuracy.

• Semantic, thematic, or attribute accuracy.

• Logical consistency.

• Completeness.

Knowledge about spatial data quality is contained in metadata, which are communicated to users through transmission of datasets by data producers.

Daily practice in the geospatial community shows numerous and constant efforts among geospatial data producers and providers to improve the quality of geospatial data, and also to make it possible for users of geospatial data to understand the effects of data quality on their work. Although geospatial data producers usually know the data they produce and their quality in detail, geospatial data users, with the exception of users within professional geospatial communities (e.g., GEOSS, etc.), mostly find information on quality too complicated and therefore either do not understand these quality details or do not even care about them. We all need to improve the understanding of geospatial data quality and its effects.

Several methodologies are known for how user requirements are turned into the quality requirements that have been developed and applied in recent decades, such as the approaches of Total Quality Management (TQM) and Workflow Management (WFM; Radwan et al., 2001). These include Quality Function Deployment (QFD), a method of Decision Support Systems (DSS) to transform user demands into design quality (Akao, 1994). QFD helps transform customer needs, using a market research technique termed Voice of the Customer (VOC) in business and information technology, into engineering characteristics and appropriate test methods for a product or service, as well as exploring how to turn user input into innovation (Ulwick, 2002). Extensive discussions on these approaches can be found in a variety of literature (e.g., Burstein and Holsapple, 2008). In this regard, Geographic Information Systems (GIS) are also described in the literature as a tool of Spatial DSS (Pick, 2005).

The next section of this paper demonstrates that specialists in geospatial data quality have good theories and techniques at their disposal. On the other hand, these theories and techniques are unfortunately not accessible to all and are still far from being sufficiently understood, especially in the user domain. The current situation in this regard is not what participants in the global geospatial data community need and expect in order to achieve quality results when using geospatial data. The size and effect of this problem will accelerate with new developments in geospatial data production principles, where the borders between geospatial data producers and data users are becoming blurred.

Various kinds of web-based user input of geospatial data, enabled by Web 2.0 technology (O’Reilly, 2005), are changing the processes of geospatial data production and use, with both good and bad consequences. Participants in the geospatial data production process are becoming both producers and users, which recently gave rise to the neologism produsers (Bruns, 2008). The terms neogeography (Turner, 2007), VGI or volunteered geographic information (Goodchild, 2007), crowdsourcing and urban sensing (McLaren, 2009), and cloud computing with its cloud service models and cloud deployment models (Mell and Grance, 2009) were coined recently and are rapidly gaining importance in the global geospatial community. Filling much of the agenda at the recent international geospatial events, it becomes obvious that the concepts of collaborative data development have very extensive effects on development practices and will revolutionize how participants in the geospatial community use and manage data. In the second half of 2009, we are witnessing the first of the established commercial producers (e.g., deCarta) of geospatial tools and services announcing support for crowd-sourced geospatial data across their product line. This is introducing new developments for companies in the geospatial tools business, for geospatial data producers, and for end users of geospatial data.

Users increasingly treat the web as an interactive environment in which the accumulation of information from individuals is as important as the distribution of information to individuals, as can be seen in various available solutions and examples such as Tele Atlas MultiNet (www.teleatlas.com), TomTom Map Share (www.tomtom.com/page/mapshare), Navteq Map Reporter (mapreporter.navteq.com), Nokia Maps (maps.nokia.com), Google Maps (maps.google.com), Bing Maps (www.bing.com/maps), Wikimapia (wikimapia.org), and OpenStreetMap (www.openstreetmap.org).

Therefore every effort to improve the systematic communication of data-quality information between data producers and data users is welcome. The authors’ contribution to spatio-temporal data-quality assessment is the concept of a simple and dynamic solution implementing the STEM (Spatio-Temporal Evaluation Matrix) and INSTANT (INdex of Spatio-Temporal ANTicipations) matrices on spatio-temporal data, explained in detail in Section 3.