We have briefly investigated how mathematical structures on data can be used to define conditions on the visualization mapping from data to displays. The first three conditions that we discussed are that D: U → V map: Thus for each of three kinds of mathematical structures on the data and display models U and V, the conditions state that D should define a correspondence between the structure of U and the structure of V. This is an interesting similarity of form between these three conditions. The fourth condition is a bit different, relating the structure of the repertoire E to the structure of a symmetry group G on the display model V. While these ideas are certainly not fully developed, they define an interesting approach to visualization. They suggest the possibility of expressing properties of human perception and visualization requirements in terms of mathematical structures, and deriving visualization mappings by mathematical analysis. We note that the problems of symbolic integration and theorem proving were once solved heuristically (i.e., solved by applying expert “rules of thumb”), and are now solved systematically. It may be possible to solve the problem of designing visualizations in a similarly systematic way.
In this paper we discuss the use of AVS5 as an interface tool to the public domain database management system POSTGRES. This work is part of ongoing research on data analysis tools for large (terabyte) data sets. Statistical analysis and browsing tools for the data sets are provided via the AVS5 and POSTGRES systems. Graphical interfaces are provided for both the feature extraction stage and the post processing stage of the system. The tools for these interfaces, rather than the statistical tools themselves, are the focus of this paper.
We describe a database management — data visualization integration based on the view concept. In general, views are descriptions of data transformations that enable more efficient interactions between user and data. Our intent is to use the view concept as the unifying link between the two systems, thereby enabling database interaction to be closely coupled with data visualization. We discuss the types of views used to realize the integration, and their implementation in our environment.
FAN (File Array Notation) is an array-oriented language for identifying data items in files for the purpose of extraction (e.g. for visualization) or modification. FAN specifications consist of These identifiers can be names or numbers. FAN is intended to be generic and could be applied to any file format, even relational databases. However netCDF is the only such format currently supported. The use of FAN can greatly simplify access to array-oriented data.
The separation of data model and presentation has been shown to be highly useful for the visualization of data. The Model-View-Controller (MVC) paradigm used by Smalltalk has been accepted as the de facto standard in this area. Over the past two years, a new software architecture has been developed based on the MVC approach. This architecture takes into account the special requirements of the C++ language, object databases (such as POET) and C++ class libraries for GUI development. It is used in commercial products, however it is not available in itself as a separate product. POETView is intended for use in internal tool production and in external project development. The development of POETView was motivated by the need for reusable user interface and data model components. As significantly more effort is generally required to develop the graphical interface of an application than the algorithms used in the model, we have tried to allow the reuse of existing components whenever possible in different contexts. However, components are not found only at the user interface level. The data model also contains reusable components. We consider reusability to be most effective when the components can be reused in binary form. These objects must therefore be able to communicate with each other at run-time which data are to be displayed. This protocol is at the heart of the POETView architecture. It allows a presentation component to obtain data from the associated data model components and to display them as desired. The protocol also manages any necessary meta-data. In this way, components can be combined at run-time, since the static or compile-time definition of possible aggregations is not necessary.
The COllaborative VIsualization and Simulation Environment (COVISE) has been designed for a distributed and collaborative scenario in a high-performance computing and networking environment. Special emphasis has been put on the handling of the data in the whole system. Its concept though is sufficiently flexible that even the (more common) scenario of desktop workstation and slower WAN-connections can benefit from the advantages of a collaborative working environment. We will give an overview of the system and its architecture, explain its advantages and present our experiences, including performance measurements.
Visualization is one of the most important activities involved in modern exploratory data analysis. Traditional database data models, in their current forms, are inadequate to satisfy the data modeling need of exploratory data analysis in general and visualization in particular. A comprehensive scientific data model is required for seamless integration of various components of a scientific database system which includes visualization, data analysis, and data management. This paper identifies the criteria of a comprehensive scientific data model and discusses the implementation aspect of such a model based on existing DBMS. Recent research in scientific data modeling and various issues raised in the workshop subgroup are also presented.
This Metadata Workgroup Report first overviews various metadata activities reported on the world wide web. The workgroup participant discussions and contributions are then summarized. A detailed discussion of metadata for scientific simulation and measurement handling has been selected. Different metadata architectures which depend on the relationship between the usage characteristics and goals of data and metadata are then described. Finally key research issues are identified. A specific emphasis of this paper is to present metadata discussion issues from the viewpoint of visualization and its integration with large scale simulation and data handling.
More and more of our customers have to deal with very large datasets like elevation data and digital roadmaps covering Europe or even the entire world, very large images e.g. from satellites or CT-Scans both medical and industrial. In this paper we describe how we tried to solve the problems of storing and especially of accessing those large datasets by developing LadMan, a Large Data Management System. We will explain LadMan's design and architecture, we will give a short overview of LadMan's API and will look at the implementation of LadMan as an important part of RUVIS, a Broadcasting Planning and Visualization System developed for German Telekom by VISTEC Software GmbH.
Data sampling is the first step in the process of scientific data analysis. This paper focuses on some mathematical aspects of data sampling. From this perspective, data are viewed as mathematical functions instead of just values. We show that continuity is the single most important quality of data (viewed as functions) which makes scientific data analysis and visualization meaningful and/or possible. By separating issues related to data functions from those related to the domains of data functions, we are able to define continuous data in two distinct contexts. This paper also provides a framework and the necessary mathematical language for the modeling and description of data.
This paper reports on user experience with Tioga, a DBMS-centric visualization tool developed at Berkeley. Based on this experience, we have designed Tioga-2 as a direct manipulation system that is more powerful and much easier to program. We present a detailed design of the revised system together with an extensive example of its application. We also give a progress report on a Tioga-2 implementation.
This paper talks about the requirements for an effective information visualization tool. The architecture of Navigational View Builder, a tool to develop visualizations of the information space of hypermedia systems is also described. We then show how the tool satisfies the requirements of an information visualization tool. To conclude the paper we discuss some of the required features of the underlying database for producing effective visualizations.
Much research has been done on interfacing databases to visualization applications, and there are many varied approaches. We describe the lessons learned during the design and development of the REINAS software (Real-time Environmental Information Network and Analysis System). We have developed visualization applications that have been in public use for several years and that visualize data from relational database engines. Our most popular tools are World Wide Web access tools. Our most sophisticated tools have novel user interfaces-Spray and CSpray. In this paper we present the interface and API (application programming interface) issues, as well as the development of the required middleware. The REINAS system is a complete data management system whose requirements and construction were driven primarily by the visualization needs, and therefore presents a unique view of how to utilize commercial relational database technologies for environmental visualization.
An important goal of visualization technology is to support the exploration and analysis of very large databases. Visualization techniques may help in database exploration by providing a comprehensive overview of the database. Pixel-oriented visualization techniques have been developed to visualize as many data items as possible on the display at one point of time. The basic idea of pixeloriented techniques is to map each data value to a colored pixel and present the data values belonging to different dimensions (attributes) in separate subwindows. In case of the query-dependent techniques, the pixels are arranged and colored according to the relevance for the query, providing a visual impression of the query result and of its relevance with respect to the query. One problem of the current query-dependent pixel-oriented visualization techniques is that their local clustering properties are insufficient. In this paper, we therefore generalize the original pixel-oriented techniques and propose new variants which retain the overall arrangement but enhance the clustering properties by using screen-filling curves locally. Different screen-filling curves (Snake, Peano-Hilbert, Morton) with different sizes (2, 4, 8, 16) may be used. We evaluate the possible variants and compare the resulting visualizations. The visualizations show that screen-filling curves clearly enhance the visual clustering of query-dependent pixel-oriented visualization techniques, but it also becomes clear that there is no significant difference between the different screen-filling curves.
One of the strangest paradoxes of the silicon era is the dichotomy between ’enjoyable’ recreational computer activities and ’mundane’ work-based computer operations. How can an activity as pointless as a computer game have so much appeal? The answer to this lies in the user interface, and not the functionality, of the program. Computer games rely heavily on an interface which is natural and enjoyable to use. We believe that an interface should appeal to the user, and to do so must capture the user's interest and imagination. To this end, we have been using high performance graphics to generate meaningful three dimensional representations for our graphical user interface. We propose new metaphors for both query construction and result representation.
The systems integration questions really extend beyond database and visualization aspects to include external data processing services such as numerical compute engines, statistical packages, neural networks, and the like. In constructing a system, several different design dimensions must be considered, such as ease of use, performance, size of data, coupling of systems, extensibility, heterogeneity, migration paths, distribution, etc., as outlined previously. In addition, in designing systems such as dataflow or object oriented modeled systems, it may be possible to connect two different procedures, or operations over data, but it may not be meaningful to do so. Efforts aimed at improving the integration of visualization and database systems are currently underway (e.g. Sequoia 2000, Aurora from Xidak, and PAGEIN), differing in scope and design goals. We hope that heightened awareness of the needs of users with data management and visualization problems will further increase interaction between developers of database and visualization systems in the near future, to address the numerous challenges that lie ahead.
This paper describes a data management strategy which implements the thesis that the design of object-oriented. scientific databases (e.g., the data structures, hashing procedures) should incorporate the domain-specific relations of objects in the database to facilitate extraction and extrapolation services required by user queries. The strategy has been used to create object-oriented databases from data sets computed by iterating over meshes, incorporating a variety of geometries and coordinate systems. One set of data structures defining the data object and the hashing maps is sufficient to manage this type of data (multiple scalar and vector quantities at nodes in an n dimensional mesh) for interactive examination on both serial and massively parallel facilities. The question raised is: can this approach which exploits the inherent logical relations of meshed data generated by simulations of fluid flows be applied to scientific data from other domains with different inherent relations?
The absence of a uniform and comprehensive representation for complex scientific data makes the adaptation of database technology to multidisciplinary research projects difficult. In this paper, we clarify the taxonomy of data representations required for scientific database systems. Then, based on our proposed scientific database environment, we present a scientific data abstraction at the conceptual level, a schema model for scientific data. This schema model allows us to store and manipulate scientific data in a uniform way independent of the implementation data model. We believe that more information has to be maintained as metadata for scientific data analysis than in statistical and commercial databases. Clearly, metadata constitutes an important part of our schema model. As part of the schema model, we provide an operational definition for metadata. This definition enables us to focus on the complex relationship between data and metadata.
