From SMW+ Semantic Enterprise Wiki
Semantic Wiki based Knowledge Management for Business Intelligence
Description
BI-Team[1] is a Dutch company which uses SMW+ for complementing its Business Intelligence framework with Knowledge Management capabilities. In today’s BI scenarios, it has become increasingly important to acquire, discover and re-use BI-related artifacts in an organized and structured manner. Another aspect that traditional BI environments often lack is the possibility of capturing, organizing and documenting BI processes and procedures. This is where the semantic wiki SMW+ comes into play. It is used within the BI framework to capture, manage and provide access to Meta- and Master data. The semantic features of SMW+ enable meaningful and structured description of these data sources thereby making them easier to discover, manage and query.
Metadata management is central in to BI and it is perceived to be one of the most vital features of BI environments[2]. Another issue that is dealt with using SMW+ in this scenario is the generation and the refinement of the ETL code.
ETL is a common procedure in BI scenarios[3] and involves three steps:
- Extraction of data from the original source data.
- Transformation of this data into a suitable representation.
- Loading of the data into a target system; typically a Data Mart or Data Warehouse.
The following figure shows the workflow that is gone through while generating ETL statements using SMW+.
The process is made up of four steps:
- A module creates semantically tagged wiki pages from the data source’s structure and content.
- A rule engine evaluates the content of the wiki and adds its conclusions to the wiki.
- A module reads the wiki and creates a XML file, describing the target Data Vault or Data Marts.
- A module reads the XML target-definition file and creates ETL code plus tagged wiki pages that describe the target and its relation to the source.
Users can refine the gained knowledge in the wiki and thus influence the ETL process in a collaborative manner at any particular stage. Due to its semantic, machine-interpretable meaning, the collected knowledge can directly flow into the automated generation of ETL statements. Everyone who is acquainted with this data can give input directly and share it with others until the optimal target definition is found.
The diagram above shows the reference architecture of BI-Team. The wiki supplements the advanced statistics and data mining components thus making it a complete and rich BI environment. With this framework, BI-Team serves customers in finance, transport and trade industry, amongst others.
Benefits
Using SMW+ in this scenario has the following benefits:
- Machine-processible, formal knowledge for ETL statements can be captured and re-used easily.
- Different data stakeholders and BI analysts can collaboratively refine knowledge and reach a consensus
- Intuitive querying and generation of reports allows monitoring and gaining of new insights.
- The meaning of data and its interrelations is not hidden in database schema; it is made explicit and preserved.
References
- ↑ http://www.bi-team.com
- ↑ Gartner: Magic Quadrant for Business Intelligence Platforms, 2009.
- ↑ Wikipedia entry about ETL
- ↑ 4.0 4.1 BI-Team: Meta- & Master Data Management (whitepaper).
