Nicolas Prat

A UML-based data warehouse design method

Data warehouses are a major component of data-driven decision support systems (DSS). They rely on multidimensional models. The latter provide decision makers with a business-oriented view to data, thereby easing data navigation and analysis via On-Line Analytical Processing (OLAP) tools. They also determine how the data are stored in the data warehouse for subsequent use, not only by OLAP tools, but also by other decision support tools. Data warehouse design is a complex task, which requires a systematic method. Few such methods have been proposed to date. This paper presents a UML-based data warehouse design method that spans the three design phases (conceptual, logical and physical). Our method comprises a set of metamodels used at each phase, as well as a set of transformations that can be semi-automated. Following our object orientation, we represent all the metamodels using UML, and illustrate the formal specification of the transformations based on OMGs Object Constraint Language (OCL). Throughout the paper, we illustrate the application of our method to a case study.

Measuring Data Believability: A Provenance Approach

Data quality is crucial for operational efficiency and sound decision making. This paper focuses on believability, a major aspect of quality, measured along three dimensions: trustworthiness, reasonableness, and temporality. We ground our approach on provenance, i.e. the origin and subsequent processing history of data. We present our provenance model and our approach for computing believability based on provenance metadata. The approach is structured into three increasingly complex building blocks: (1) definition of metrics for assessing the believability of data sources, (2) definition of metrics for assessing the believability of data resulting from one process run and (3) assessment of believability based on all the sources and processing history of data. We illustrate our approach with a scenario based on Internet data. To our knowledge, this is the first work to develop a precise approach to measuring data believability and making explicit use of provenance-based measurements.

Multidimensional Schemas Quality: Assessing and Balancing Analyzability and Simplicity

A data warehouse is a database focused on decision making. Decision makers typically access data warehouses through OLAP tools, based on a multidimensional representation of data. In the past, the key issue of data warehouse quality has often been centered on data quality. However, since OLAP tool users directly access multidimensional schemas, multidimensional schema quality evaluation is also crucial. This paper focuses on the quality of multidimensional schemas, more specifically on the analyzability and simplicity criteria. We present the underlying multidimensional model and address the problem of measuring and finding the right balance between analyzability and simplicity of multidimensional schemas. Analyzability and simplicity are assessed using quality metrics which are described and illustrated based on a case study. The main objective of our approach is to provide the data warehouse designer with precise measures to support him in the choice among several alternative multidimensional schemas.

Dimension Hierarchies Design from UML Generalizations and Aggregations

Data for decision-making applications are based on dimensions, such as time, customer, and product. These dimensions are naturally related by hierarchies. Hierarchies are crucial to multidimensional modeling. Defining hierarchies using star or snowflake schemas can be misleading, since they are not explicitly well-modeled. However, deriving them from conceptual UML or ER schemas is a non-trivial task since they have no direct equivalent in conceptual models. This paper focuses on the definition of multidimensional hierarchies. We present and illustrate rules for defining multidimensional hierarchies from UML schemas, especially based on aggregation and generalization hierarchies. The definition of hierarchies is part of a data warehouse design method based on the three usual modeling levels : conceptual, logical, and physical. The conceptual schema is based on the UML notation. The logical schema is represented using a unified pivot multidimensional model. The physical schema depends on the target ROLAP or MOLAP tool.

A Taxonomy of Evaluation Methods for Information Systems Artifacts

Abstract Artifacts, such as software systems, pervade organizations and society. In the field of information systems (IS) they form the core of research. The evaluation of IS artifacts thus represents a major issue. Although IS research paradigms are increasingly intertwined, building and evaluating artifacts has traditionally been the purview of design science research (DSR). DSR in IS has not reached maturity yet. This is particularly true of artifact evaluation. This paper investigates the “what” and the “how” of IS artifact evaluation: what are the objects and criteria of evaluation, the methods for evaluating the criteria, and the relationships between the “what” and the “how” of evaluation? To answer these questions, we develop a taxonomy of evaluation methods for IS artifacts. With this taxonomy, we analyze IS artifact evaluation practice, as reflected by ten years of DSR publications in the basket of journals of the Association for Information Systems (AIS). This research brings to light important relationships between the dimensions of IS artifact evaluation, and identifies seven typical evaluation patterns: demonstration; simulation- and metric-based benchmarking of artifacts; practice-based evaluation of effectiveness; simulation- and metric-based absolute evaluation of artifacts; practice-based evaluation of usefulness or ease of use; laboratory, student-based evaluation of usefulness; and algorithmic complexity analysis. This study also reveals a focus of artifact evaluation practice on a few criteria. Beyond immediate usefulness, IS researchers are urged to investigate ways of evaluating the long-term organizational impact and the societal impact of artifacts.

Combining objects with rules to represent aggregation knowledge in data warehouse and OLAP systems

Data warehouses are based on multidimensional modeling. Using On-Line Analytical Processing (OLAP) tools, decision makers navigate through and analyze multidimensional data. Typically, users need to analyze data at different aggregation levels (using roll-up and drill-down functions). Therefore, aggregation knowledge should be adequately represented in conceptual multidimensional models, and mapped in subsequent logical and physical models. However, current conceptual multidimensional models poorly represent aggregation knowledge, which (1) has a complex structure and dynamics and (2) is highly contextual. In order to account for the characteristics of this knowledge, we propose to represent it with objects (UML class diagrams) and rules in the Production Rule Representation language (PRR). Static aggregation knowledge is represented in the class diagrams, while rules represent the dynamics (i.e. how aggregation may be performed depending on context). We present the class diagrams, and a typology and examples of associated rules. We argue that this representation of aggregation knowledge enables an early modeling of user requirements in a data warehouse project. A prototype has been developed based on the Java Expert System Shell (Jess).

Multidimensional models meet the semantic web: defining and reasoning on OWL-DL ontologies for OLAP

Data warehouses use a multidimensional model. Based on this model, OLAP cubes enable users to analyze data. For correct OLAP analysis, multidimensional models should be checked. In particular, these models should ensure summarizability. Checking multidimensional models and their summarizability is complex and error-prone. To perform this task, formal reasoning is appropriate. In this paper, we propose and illustrate an approach to represent a multidimensional model as an OWL-DL ontology, and reason on this ontology to check the multidimensional model and its summarizability. Beyond the reasoning capabilities of description logic, representing multidimensional models as OWL-DL ontologies is a means to move multidimensional modeling to the semantic Web. To illustrate this, we investigate the complementarities between our approach and the RDF Data Cube vocabulary, and suggest how they could be combined.

Transforming multidimensional models into OWL-DL ontologies

Business intelligence is based on data warehouses. Data warehouses use a multidimensional model, which represents relevant facts and their measures according to different dimensions. Based on this model, OLAP cubes may be defined, enabling decision makers to analyze and synthesize data. Ontologies (and, more specifically, OWL ontologies) are a key component of the semantic Web. This paper proposes an approach to represent multidimensional models as OWL-DL ontologies. To this end, it presents the multidimensional metamodel, the concepts of OWL-DL, and transformation rules for mapping a multidimensional model into and OWL-DL ontology. It then illustrates application to a case study with a simplified example of a spatiotemporal data warehouse. The transformation rules are refined to deal with spatiotemporal data warehouses, applied step by step, and the resulting ontology is implemented in the Protégé ontology tool. As illustrated by the case study, our approach enables better formalization and inferencing, thanks to OWL-DL. The ontology may also be used to represent OLAP cubes on the semantic Web (with RDF), by defining these cubes as instances of the OWL-DL multidimensional ontology.

A Context-Aware Architecture for Mobile Knowledge Management

In many professional activities (e.g., medical diagnosis, construction, or sales), the ability to retrieve and store knowledge in a mobile situation is crucial. This need, together with the progress of mobile devices, has led to the emergence of mobile knowledge management. The mobile situation imposes specific constraints on traditional knowledge management activities (e.g., knowledge retrieval or presentation). Therefore, a key research question for mobile knowledge management is how context should be taken into account.

Business Intelligence and Big Data in the Cloud: Opportunities for Design-Science Researchers

Cloud computing and big data offer new opportunities for business intelligence (BI) and analytics. However, traditional techniques, models, and methods must be redefined to provide decision makers with service of data analysis through the cloud and from big data. This situation creates opportunities for research and more specifically for design-science research. In this paper, we propose a typology of artifacts potentially produced by researchers in design science. Then, we analyze the state of the art through this typology. Finally, we use the typology to sketch opportunities of new research to improve BI and analytics capabilities in the cloud and from big data.