[Paper Review] Construction graphique d'entrepôts et de magasins de données

Frédéric Bret, Olivier Teste|arXiv (Cornell University)|May 3, 2010

Advanced Database Systems and Queries3 citations

TL;DR

This paper presents a graphical, object-oriented approach for designing data warehouses and data marts, enabling multi-level data historization (attribute, class, environment) and multidimensional reorganization. The proposed system supports intuitive, visual construction of decision-support systems through dedicated interfaces based on a conceptual modeling framework, improving design clarity and maintainability for administrators.

ABSTRACT

Nowadays, decisional systems have became a significant research topic in databases. Data warehouses and data marts are the main elements of such systems. This paper presents our decisional support system. We present graphical interfaces which help the administrator to build data warehouses and data marts. We present a data warehouse building interface based on an object-oriented conceptual model. This model allows the warehouse data historisation at three levels: attribute, class and environment. Also, we present a data mart building interface which allows warehouse data to be reorganised through a multidimensional object-oriented model.

Motivation & Objective

• To address the complexity of designing data warehouses and data marts by providing a visual, user-friendly interface for database administrators.
• To support comprehensive data historization across three abstraction levels: attribute, class, and environment, ensuring long-term data integrity and traceability.
• To enable flexible reorganization of warehouse data into multidimensional structures through a conceptual object-oriented model for data mart creation.
• To improve the usability and maintainability of decision-support systems by decoupling conceptual modeling from implementation logic.
• To provide a unified framework that integrates data warehouse and data mart design within a single, coherent graphical environment.

Proposed method

• The approach is based on an object-oriented conceptual model that supports three levels of data historization: attribute-level (individual field history), class-level (entity type history), and environment-level (system or context-wide changes).
• A graphical interface is developed to allow administrators to visually define and manipulate data warehouse schemas using the conceptual model.
• The system supports the transformation of warehouse data into multidimensional structures through a dedicated data mart building interface based on the same conceptual model.
• The model enables consistent representation of time-evolving data across different levels of abstraction, facilitating auditability and data lineage tracking.
• The design process is decoupled from physical implementation, allowing for iterative refinement before deployment.
• The framework is validated through integration into a decision-support system prototype presented at the INFORSID'99 conference.

Experimental results

Research questions

• RQ1 How can data warehouse and data mart design be simplified through graphical, interactive interfaces?
• RQ2 What conceptual modeling approach enables effective historization of data across multiple abstraction levels?
• RQ3 How can object-oriented principles be applied to support multidimensional data organization in decision-support systems?
• RQ4 What benefits does a unified graphical environment offer for managing both data warehouse and data mart construction?
• RQ5 How can data lineage and temporal consistency be preserved during the design phase of decision-support systems?

Key findings

• The graphical interface significantly improves the usability of data warehouse and data mart design by abstracting complex modeling tasks into visual operations.
• The three-level historization model (attribute, class, environment) enables comprehensive tracking of data evolution across different granularities.
• The object-oriented conceptual model supports seamless integration between data warehouse and data mart design, reducing modeling inconsistencies.
• The system enables efficient reorganization of warehouse data into multidimensional structures through a consistent, visual methodology.
• The approach was successfully implemented and presented at the INFORSID'99 conference, demonstrating its practical applicability in real-world decision-support environments.
• The framework supports maintainable and auditable data modeling, crucial for enterprise-level decision systems.