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

Multidimensional modeling and analysis of large and complex watercourse data: an OLAP-based solution

“…The classical relational OLAP architecture is composed of four tiers (Boulil et al, 2013b;Malinowski and Zimányi, 2008): ETL, data storage tier, OLAP server and OLAP client.…”

“…used by water quality practitioners are not suited to manage and to analyze large volumes of information. Recently, some studies have shown the ease and power of using Data Warehouse (DW) and On-Line Analytical Processing (OLAP) technologies to store and to analyze environmental data (Alexandru et al, 2010;Boulil et al, 2013b;McGuire et al, 2008;Pinet and Schneider, 2010). DWs are databases dedicated to the integration and storage of large volumes of data to support the decision processes of organizations (Inmon, 2005).…”

“…DW and OLAP data cubes are generally implemented in relational platforms consisting of four tiers (Boulil et al, 2013b;Malinowski and Zimányi, 2008): the ETL that integrates data from data sources into the DW, the DW (a relational database that stores the finest data), the OLAP server that calculates the data cubes from the DW data and the OLAP client that displays the data cube information using tables and different types of statistical diagrams (pie charts, histograms, etc.) and reports in different export formats.…”

“…Using our framework (Boulil et al, 2013a) that we extend here by complex aggregate functions (e.g., a generic function to calculate all percentiles), we show how to define complex analysis indicators by using these complex functions and by introducing additional analysis dimensions to allow their calculation and also for information rendering purposes. In Boulil et al (2013aBoulil et al ( , 2013b, we have shown how to define complex indicators having multiple but only simple aggregate functions. In addition, we propose two strategies to address the heterogeneity of measurement units (one of the main summarizability semantic problems) by (i) transforming source data at the ETL tier, and (ii) by introducing an additional analysis dimension at the OLAP server tier.…”

“…Semantic and structural summarizability conditions grant the accuracy and the correctness of indicator values if we assume a good DW data quality (please see Boulil et al (2012) for more details about the quality of OLAP analysis indicators). Finally, this paper constitutes a second application in the environmental domain of our framework for data cubes (Boulil et al, 2013a); a first experiment of this framework in agriculture is presented in (Boulil et al, 2013b). Our framework considerably reduces the development times and effort by automating most of the implementation tasks.…”