Business Intelligence Glossary
Business Intelligence has a very broad meaning. The system is mostly defined as „integrated set of tools, technologies and software products used to collect data from different diffused sources, to integrate, analyze and share them”. Thanks to the ability to connect and analyze data including the whole area of a company activity (see more: ETL Process) according to the accepted business operations they allow to manage the knowledge in an organization, symulate and forecast the consequences of operations and improve decision-making processes thanks to interactive reports and management dashboard. In ERP systems the way of data storage is optimized to manage current operations. However, in case of analitical queries it can cause significant elongation of processing time. The generated results are most often presented in a two-dimensional table which in case of advanced analysis containing bigger amount of variables may not be sufficient for accurate analysis. See more: What benefits does the implementation of Business Intelligence give?
ETL process (Extraction, Transformation and Loading) integrates data which come from different sources to ensure high quality and coherent data. This process is central to construct and maintain data warehouse. Only properly integrated and transformed data will be useful in business analysis.
See more: Which data sources can you connect with Business Intelligence system?
See more: Is it possible to successively add new source systems to Business Intelligence system?
Dashboards presents main data concerning the organization. They have a lot of interactive graphics, i.a.: sliders, drop-down menu, widgets, which allow to develop possibility of data analysis. They make possible e.g. analysis of KPI (Key Performance Indicators).
These tools allow to analyse the functioning of a company, to increase managers decisions pertinence, they facilitate management of information flow. They also allow to monitor the company results. Thanks to all of these the efficiency of the whole organization increases.
Thanks to the KPI (Key Performance Indicators) a user can check the level of business processes implementation (e.g. sale) and evaluate the effectiveness of their realization (e.g. cost, time, quality). KPI enable the efficient analysis of the results, thereby it is possible to quickly react to the emerging problems (e.g. fall in sales) or monitor the efectiveness of the introduced changes.
KPI are often presented as indicators or arrows, what you can see below.
DWH (Data Warehouse) is a base where the data is stored, in a way that makes fast and efficient analysis possible. The results can be used in decision-making process.
Market Basket Analysis imitates human brain work – it associates components, unites them and finds mutual relations. This is the base of a unique technology, which makes doing all calculations in real time possible, in the same moment when a user is clicking. Searching for some information is as simple as in the popular browser Google. It is possible to make the analysis very fast, without need to reduce it to the predefined query paths only, always seeing all the dependences between the data. While building the classic BI system it is necessary to define relations between the data, especially in the hierarchy.
Creating a data model, it is almost always indispensable to analize them step by step to consider and configure all the relations. In the next step one should add essentials hierarchies, which are caused by the earlier analysis. They give the opportunity to drill. Every new need brings changes into the model and causes manual addition of relations and hierarchies. In Associative Logic all the data is automatically conected with each others. If any relation is unnecessary or wrong, it is removed. It speeds up the work during the unite data modeling to introduce them in reports.
As you can see on scheme, between selected objects all possible relations are defined. Thanks to this – choosing e.g. one of the administrative disctricts, the user automatically achieves information about its regions, sold products and the data of sales representatives working in this area.
OLAP (Online Analytical Processing) – a class of IT systems to suport tactical and strategic organization activities, in opposite to transactional systems which are supporting current operating activity. There are different types of OLAP in use. The most popular implementation of this type of system is data warehouse, often called relational OLAP – ROLAP. Data warehouse is based on the same solution as transactional database of more denormalize structure – like e.g. a star or a snowflake. Intentional denormalization allows to efficiently query big groups of data, e.g. to make reports. Multidimensional cube is also the OLAP type of solution. Mostly it is another layer of abstraction set on data warehouse, where the data is storaged in files of special multidimensional structures. MOLAP has its own engine of processing, which is an for intuitive and fast way of quering data in mentioned files. Earlier calculated aggregations are also storaged in multidimensional structures. Thanks to the database engine does not have to process the data on high level of details. In fact OLAP is often identified with multidimensional analytic cube. A lot of tools like SQL Server let us create a hybrid version between relational and multidimensional OLAP.
In-memory is one of leading technologies in IT development systems. Until now traditional IT tools, connected with databases and Business Intelligence, were based on downloading the data from files located on the hard drive. During the years speed of readout has increased, thanks to indexes and introduction SSD memories. In-memory greatly increased effectiveness. This disk is used to store, not to process, data. During reading from the disk, all data is loaded to the memory. Then data processing is going on. RAM memory is much faster than all availables hard drives. It is because of the overhead time elimination, what cause that data is generated faster. Using tools based on the In-memory technology, the use of RAM memory increases, so machines mus be equipped with an adequate amount of memory. This technology is used e.g. in Qlik View, Microsoft Power Pivot or SAP HANA.
The basic difference between these technologies is a place where the data is storaged and processed. The traditional approach claims that there exists the central data repository, where client tools are strictly connected with the server repository.
Self-Service BI (often called as Personal BI) is something different. It doesn’t need central object in its architecture, but the role of client tools is enhaced. Using these tools it is possible to create analytical models. The user can also modify existing solutions or even create them, without need to involve a lot of people. Self-Service BI, in contrast to the traditional approach, doesn’t require expensive investment in machines. But it also doesn’t guarantee “the only one version of the truth” and easy documents sharing.
The existence of several different solutions of the same business area can cause a real implementational problem. However, both solutions may exist simultaneously giving an opportunity to eliminate defects and emphasize advantages. One of the hybrid solutions is rearing Pover Pivot model on Sharepoint server. Thereby, a lot of final users can use a specific model, modify it and share with others.
Presently a lot of software providers allow to create Self-Service BI solutions, e.g. Microsoft Self-Service BI or QlikView.
Self-Service BI conception in Microsoft technology is based on MS Excel.
Tradicional Microsoft Corporate BI conception.