The main objective of Big data is to meet business competitiveness and facilitate decision-making. To achieve this goal, the implementation of a governance plan is necessary. In this article, we propose a big data governance evaluation approach based on the fact that big data governance is seen as an artifact having the characteristics of a system. We develop a multi-criteria hierarchy organized according to the main dimensions of systems theory. The approach is illustrated on a practical case.

International audience

Most NoSQL databases are schemaless. Although they offer some flexibility, they do not have any knowledge of the database schema, losing the benefits provided by these schemas. It is generally accepted that data modelling can have an impact on performance, consistency, usability, and maintainability. We argue that NoSQL databases need data models that ensure the proper storage and the relevant querying of the data. This paper seeks to present and illustrate an MDA-based approach, allowing us to achieve a reverse engineering of NoSQL property graph databases into an Extended Entity-Relationship schema. The approach is applied to the case of Neo4j graph database. We present an illustrative scenario and evaluate the reverse engineering approach.

The recent buzz around open data highlighted the crucial problem of anonymization in the context of data publishing. Many research efforts were devoted to the definition of techniques performing such an anonymization. However the selection of the most relevant technique and the adequate algorithm is complex. Successful decision depends firstly on the ability of data publishers to understand the anonymization techniques and their associated algorithms. In this paper, we focus on the choice of an algorithm among the different ones implementing one of the anonymization techniques, namely generalization. Through an abstraction process presented in this paper, we provide data publishers with simplified descriptions for the generalization technique and its algorithms. These descriptions facilitate the understanding of the algorithms by data publishers having low programing skills. We present also some other use cases of these abstractions as well as an experimentation conducted to validate them.

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.

Conceptual models are of prime importance to ensure a high level of quality in designing information systems. It has been witnessed that the majority of information systems (IS) change requests result due to deficient functionalities in the information systems. Therefore, a good analysis and design method should guarantee that conceptual models are correct and complete and easy to understand, as they are the communicating mediator between the users and the development team. Similarly, if models are complex then their extension or the incorporation of missing requirements gets very difficult for the designers. Our approach evaluates the conceptual models on multiple levels of granularity in addition to providing the corrective actions or transformations for improvement. We propose quality patterns to help the non-expert users in evaluating their models with respect to their quality goal. This paper also illustrates our approach by describing an evaluation and improvement process using a case study.

Frequently the behaviour of an information system is functionally correct, but it does not meet some quality criteria, such as completeness, consistency, and usability. One way to enhance the capability of an information system is to consider its conceptual model quality as well as its functional behaviour. Conceptual model quality can be defined as a set of perceivable characteristics expressed with quantifiable parameters. The aim of this empirical investigation is to evaluate the quality of different potential conceptual models of the same universe of discourse by different Information Systems (IS) stakeholders. This paper describes: a) a set of quality factors (clarity, simplicity, expressiveness, minimality) applied to different versions of entity-relationship (ER) conceptual schemas, b) an approach enabling a comprehensive comparison of the conceptual schemas, c) an experimentation leading to the evaluation of the same schemas by IS stakeholders such as designers, end-users, and students, based on a sample of about 120 observations using different statistical methods. First results indicate that there exists a strong independence between the IS stakeholders and the quality factors used. A second result reveals a significant difference between groups of respondents in their ways to perceive conceptual schemaspsila quality. Based on our experiment, we are able to identify quality factors relevant to different groups of stakeholders, depending on several dimensions, such as their professional experience, and/or their specialization degree.