TestmanagementChanging the world one bug after amother…

Business Intelligence (BI) means the methods and procedures for semantic analysis of data in electronic form. The term intelligence in the German language is understood as information.

Its goal is the acquisition of knowledge to enhance operative and strategic decision making.

Under ideal conditions, a
Business Intelligence Competence Center (BICC) exists which is in permanent communication with all departments, and through which Business Intelligence Business Benefits can be created (BI-Requirement-Management).

A Data Warehouse is basically structured as follows:
  • Delivery through data sources, for example: ERP, CRM, SAP and Flatfiles
  • Storage and activation in DWH (Staging, Data Mart, Meta-Daten) through the ETL-Process (Extraction, Transformation and Loading)
  • Reporting of information systems, analytical applications and Data Mining to the user.

Already in the ETL-Process a lot of questions will arise which can be relevant to testing:

  • Identification of defective data through proving plausibility.
  • Identification of defective data from the previous systems.
  • Technical sources of mistakes which are not transparent.
  • Data from heterogenes sources which are not attuned and are not in the expected format of the DWH.
  • Possibly aggregations are missing so that the desired output cannot be produced.
Department-specific requirements will be stored in Data Marts. These are mostly pre aggregated and therefore well functioning.
OLAP (Online Analytical Processing), the increasing accumulation of data will be provided to the management.
Here multidimensional views on aggregations from
operative data in OLAP Cubes play a role.

Auswahl einiger Grundoperationen:

Cut to a slice


Part of a cube

Rotation and dimension

Drill Down
Detailed Values breakdown

Drill Up
Condensed to a higher level

The information is provided in the form of reports through the front-ends. Here are possible variations.
One of the main problems, however, is
by-pass-reporting. This usually occurs if reports based on parallel data selections are created out of the operative system.

Thereby a lot of negative effects can arise:
  • no reporting safety
  • loss of quality
  • lack of transparency
  • lack of trust
  • lack of acceptance

Among many of the new terms created in Information Technology (IT), the following are essential:

Big Data describes data quantities which are too large, too complex or too quickly changing to produce using classic methods of IT.
Often the complexities of technology are described with the term Big Data which is used for the collecting and evaluating such quantities of data.

Hadoop Distributed File System (HDFS) is a highly available, efficient data system for storage of very large qualities of data in the filing systems of many computers.
HDFS supports data systems with several hundred million files.
Cloudera and Hortonworks are two well known providers of Hadoop.
The expectations of a well functioning Data Warehouse System are extremely high.
Often expectations and reality fall far apart.
We try to close this gap.
© 2018 Holger Mayer Consulting HMC2