Kamal Boulil

Un modèle UML et des contraintes OCL pour les entrepôts de données spatiales: De la représentation conceptuelle à l'implémentation

Spatial Data Warehouses (SDW) and Spatial OLAP (SOLAP) systems represent an effective solution to perform spatial analysis on geographical phenomena. However, the quality of such analysis heavily depends on the quality of stored data and how these data are explored: how the different indicators are computed (What aggregate functions are applied to summarize the measures and in what order these functions are applied?). In this context, a number of studies have been attempted to address the issues of data quality in SDW by using Integrity Constraints (IC). In this paper, motivated by the lack of Model Driven Architecture (MDA)-based implementations, we propose a conceptual framework based on two new classifications to ease identification and implementation of SDW IC. Moreover, following an MDA approach, we propose the MDA-based modeling of most IC categories using the UML (Unified Modeling Language) and OCL (Object Constraint Language) standard languages; and show the automatic implementation of some IC classes using an MDA-based code generator, called Spatial OCL2SQL.

Guaranteeing the quality of multidimensional analysis in data warehouses of simulation results: application to pesticide transfer data produced by the MACRO Model

Currently, the vital impact of environmental pollution on economic, social and health dimensions has been recognized. The need for theoretical and implementation frameworks for the acquisition, modeling and analysis of environmental data as well as tools to conceive and validate scenarios is becoming increasingly important. For these reasons, different environmental simulation models have been developed. Researchers and stakeholders need efficient tools to store, display, compare and analyze data that are produced by simulation models. One common way to manage simulation results is to use text files; however, text files make it difficult to explore the data. Spreadsheet tools (e.g., OpenOffice, MS Excel) can help to display and analyze model results, but they are not suitable for very large volumes of information. Recently, some studies have shown the feasibility of using Data Warehouse (DW) and On-Line Analytical Processing (OLAP) technologies to store model results and to facilitate model visualization, analysis and comparisons. These technologies allow model users to easily produce graphical reports and charts. In this paper, we address the analysis of pesticide transfer simulation results by warehousing and OLAPing data, for which the data results from the MACRO simulation model. This model simulates hydrological transfers of pesticides at the plot scale. We demonstrate how the simulation results can be managed using DW technologies. We also demonstrate how the use of integrity constraints can improve OLAP analysis. These constraints are used to maintain the quality of the warehoused data as well as to maintain the aggregations and queries, which will lead to better analysis, conclusions and decisions.

Towards the definition of spatial data warehouses integrity constraints with spatial OCL

Spatial Data Warehouses (SDWs) and Spatial OLAP systems are an effective solution to perform multidimensional spatial analysis on geographical data. However, the quality of such analysis heavily depends on the quality of stored data. Due to this, some works have been attempted to address the issues of data quality of SDWs using Integrity Constraints (ICs). In this paper, we propose a conceptual framework for SDW ICs based on two new classifications. Moreover, we describe the Model-Driven Architecture (MDA)-based specification and implementation of some IC classes using the Object Constraint Language (OCL) and an MDA-based code generator, called Spatial OCL2SQL

Definition and analysis of new agricultural farm energetic indicators using spatial OLAP

Agricultural energy consumption is an important environmental and social issue. Several diagnoses have been proposed to define indicators for analyzing energy consumption at large scale of agricultural farm activities (year, farm, family of production, etc.). However, to define ad-hoc environmental energetic policies to better monitor and control energy consumption, new indicators at a most detailed scale are needed. Moreover, by defining detailed scale indicators, large quantities of geo-referenced data need to be collected to feed these energetic diagnoses. This huge volume of data represents another important limitation of systems that implement these diagnoses because they are usually based on classical data storage systems (such as spreadsheet tools and Database Management Systems). These systems do not allow for interactive analysis at different granularities/scales of huge volumes of data and do not provide any cartographic representation. By contrast, Spatial OLAP (SOLAP) and spatial data warehouse (SDW) systems allow for the analysis of huge volumes of geo-referenced data by providing aggregated numerical values visualized by means of interactive tabular, graphical and cartographic displays. Thus, in this paper, we (i) propose new appropriate indicators to analyze agricultural farm energy performance at a detailed scale and (ii) show how SDW and SOLAP technologies can be used to represent, store and analyze these indicators by simultaneously producing expressive reports.

Spatial OLAP integrity constraints: From UML-based specification to automatic implementation: Application to energetic data in agriculture

Spatial OnLine Analytical Processing systems (SOLAP) are Business Intelligence technologies allowing efficient and interactive analysis of large spatial data cubes. In this type of systems, the correctness of analysis depends on the warehoused data quality, how aggregations are performed and how data cubes are explored. In this paper, we study quality control techniques (based on integrity constraints) related to exploration of spatial data cubes. We extend our Unified Modeling Language (UML) framework previously proposed with a UML profile allowing the conceptual design of several classes of exploration integrity constraints. We also propose a tool for their automatic implementation. We validate our proposal in the context of a real case study concerning the analysis of energetic farm indicators.

Using OCL to Model Constraints in Data Warehouses

Recent research works propose using Object-Oriented OO approaches, such as UML to model data warehouses. This paper overviews these recent OO techniques, describing the facts and different analysis dimensions of the data. The authors propose a tutorial of the Object Constraint Language OCL and show how this language can be used to specify constraints in OO-based models of data warehouses. Previously, OCL has been only applied to describe constraints in software applications and transactional databases. As such, the authors demonstrate in this paper how to use OCL to represent the different types of data warehouse constraints. This paper helps researchers working in the fields of business intelligence and decision support systems, who wish to learn about the major possibilities that OCL offer in the context of data warehouses. The authors also provide general information about the possible types of implementation of multi-dimensional models and their constraints.

A Holistic Approach to OLAP Sessions Composition: The Falseto Experience

OLAP is the main paradigm for flexible and effective exploration of multidimensional cubes in data warehouses. During an OLAP session the user analyzes the results of a query and determines a new query that will give her a better understanding of information. Given the huge size of the data space, this exploration process is often tedious and may leave the user disoriented and frustrated. This paper presents an OLAP tool named Falseto (Former AnalyticaL Sessions for lEss Tedious Olap), that is meant to assist query and session composition, by letting the user summarize, browse, query, and reuse former analytical sessions. Falsetos implementation on top of a formal framework is detailed. We also report the experiments we run to obtain and analyze real OLAP sessions and assess Falseto with them. Finally, we discuss how Falseto can be seen as a starting point for bridging OLAP with exploratory search, a search paradigm centered on the user and the evolution of her knowledge.