OLAP technologies

Lecture

The main distinguishing feature of OLAP technology is that operational data analysis software tools are designed for the user to communicate with the problem rather than with the computer, and are aimed at use not by information technology specialists or expert statisticians, but by professionals in the field of management — analyst specialists, middle managers, and senior management.

Note!

The basic principles of online analytical processing (OLAP) technology:

• multidimensional conceptual representation of data, intuitive data manipulation, batch data extraction, its accessibility and level of detail;
• multi-user support based on a «client-server» architecture, transparent access to data from the desktop;
• processing of non-formalized data, separate storage of source data and the results of operational processing, processing and exclusion of missing values;
• automatic configuration of the physical level of data extraction, flexibility of report generation and standard report performance.

The OLAP service is a tool for analyzing large volumes of data in real time. A multidimensional OLAP cube and a system of corresponding mathematical algorithms for statistical processing make it possible to analyze data of practically any complexity over any time intervals. All work with the OLAP system takes place in terms of the subject domain and makes it possible to build statistically sound models of a business situation. By interacting with the OLAP system, a manager can quickly review information of interest, obtain arbitrary data slices, and perform analytical operations of drill-down, roll-up, cross-tabulation, and comparison over time simultaneously across many parameters.

Having at their disposal flexible mechanisms for data manipulation and visual display of information, the user first examines from various angles the data that may (or may not) be related to the problem being solved. He then compares various indicators with one another, trying to identify hidden interrelations. He may examine the data more closely by detailing it, for example by breaking it down by time, by region, or by client. The user may do the opposite — generalize the representation of the information even further by removing distracting details. After this, using a module for statistical estimation and simulation modeling, several scenarios for the development of events are built, from which the most acceptable option is chosen.

Note!

Standard methods of analysis determined by the nature of OLAP technology:

• factor (structural) analysis, for example, analysis of the sales structure to identify the most important components in the breakdown of interest;
• dynamics analysis (regression analysis — trend identification), for example, identifying trends and seasonal fluctuations with a clear graphical display of the dynamics;
• dependency analysis (correlation analysis), for example, comparing the sales volumes of different goods over time to identify the necessary assortment — «the basket»;

• comparison (comparative analysis), for example, comparing sales results over time or for a given period, or for a given group of goods;
• study of probability distributions and confidence intervals (variance analysis). Used, for example, for forecasting and risk assessment.

These types of analysis do not exhaust the capabilities of OLAP. For example, by applying statistical analysis functions — variance, mean deviation, higher-order moments — to calculate intermediate and final totals, one can obtain the most diverse kinds of analytical reports.