Analysis Groups for Semantic Layer Data Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems fail to efficiently represent and navigate sales data tracked at different levels of granularity, leading to data inconsistencies and discrepancies due to the lack of semantic relationships between analysis levels, making it difficult to evaluate the reliability of retrieved data.
Innovation Solution
The implementation of an abstraction layer with analysis group objects that link sales measures to their respective dimensions, allowing for different levels of analysis within a single semantic measure, and enabling navigation between these levels, reducing the need for multiple unrelated representations and providing meaningful data evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If sales data is tracked at different levels of granularity by different departments, then detailed analysis capabilities are improved, but data inconsistencies and reliability evaluation difficulties arise
Solution Approach 1:
The patent segments the sales data system into multiple analysis levels (corporate level, line of business level, product level) with each level having its own data table and measure objects. This segmentation allows each department to track data at their required granularity while the system maintains semantic relationships between levels through the abstraction layer, resolving the contradiction between detailed analysis capability and data consistency.
2Ease of operation
If multiple independent measures are created for each level of granularity, then navigation between levels is simplified, but the system complexity and number of objects increase
Solution Approach 1:
The patent introduces an abstraction layer with semantic measure objects that act as intermediaries between the physical database tables at different granularity levels. Instead of creating multiple independent measures, a single semantic measure object can reference multiple physical tables through the abstraction layer, enabling navigation between levels while reducing system complexity by eliminating redundant measure definitions.
Data Source
AI summary
A system may include a database of physical data tables including stored data, and an abstraction layer associated with the stored data. The abstraction layer may include a measure object associated with a measure, a plurality of dimension objects associated with respective dimensions, a first analysis group object linked to the measure object, to a first one or more of the plurality of dimension objects, and to a first portion of the stored data associating the measure with respective dimensions of the first one or more of the plurality of dimension objects, and a second analysis group object linked to the measure object, to a second one or more of the plurality of dimension objects, and to a second portion of the stored data associating the measure with respective dimensions of the first one or more of the plurality of dimension objects and of the second one or more of the plurality of dimension objects.