Classification of the products business intelligence

The body of BI tools is divided into corporate BI sets and BI platforms. Means of generation of requests and reports are to a large extent absorbed and substituted with corporate BI sets. The multidimensional OLAP mechanisms or servers and also the relational OLAP mechanisms are BI tools and infrastructure for BI platforms.

The majority of BI tools are used by end users to access, the analysis and reports generation by data most of which often are located in storage, data marts or operational warehouses of data. Application developers use BI platforms for creation and implementation of BI-applications which are not considered as BI tools. An example of a BI-application is the executive information system of EIS.

Instruments of generation of requests and reports

Generators of requests and reports — typically "desktop" tools providing to users access to databases, making some analysis and creating reports. Requests can be as unplanned (ad hoc), and to have procedural character. There are systems of reports generation (as a rule, server) which support procedural requests and reports. Desktop generators of requests and reports are expanded also some facilitated opportunities of OLAP. The developed tools of this category integrate in themselves possibilities of package generation of procedural reports and desktop generators of requests, mailings of reports and their operational updating, forming the so-called corporate reporting (corporate reporting). The server of reports, means of mailing, the publication of reports on Web, the mechanism of the notification on events or deviations (alerts) enter its arsenal. Characteristic representatives — Crystal Reports, Cognos Impromptu and Actuate e.Reporting Suite.

OLAP or the developed analytical tools

The OLAP tools are analytical tools which were originally based on the multidimensional databases (MD). MBD are the databases designed especially for support of the analysis of quantitative data with a set of measurements, contain data in "purely" multidimensional form. The majority of applications include measurement of time, another dimensions can concern geography, organization units, clients, products, etc. OLAP allows to organize measurements in the form of hierarchy. Data are presented in the form of hyper cubes (cubes) — logical and physical models of the indicators which are collectively using measurements and also hierarchies in these measurements. Some data are previously aggregated in a DB, others are calculated "on the fly".

Means of OLAP allow to investigate data on different measurements. Users can select what indicators to analyze, what measurements and how to display in a crosstable, to exchange lines and the pivoting columns, then to make cuts and cuttings ("slice&dice") to concentrate on a certain combination of dimensions. It is possible to change detail of data, moving on levels using detailing and enlargement of "drill down/roll up" and also cross-detailing of "drill across" through another dimensions.

For support of MBD the OLAP servers optimized for the multidimensional analysis and delivered with analytical opportunities is used. They provide good performance, but usually require a lot of time for loading and the MBD expansion. Are delivered with a possibility of "reach-through", allowing to pass from units to parts in relational DB. The classical OLAP server — Hyperion Essbase Server.

Today relational DBMS are applied to emulation of MBD and support the multidimensional analysis. OLAP for relational DB (ROLAP) has advantage on scalability and flexibility, but loses on performance to multidimensional OLAP (MOLAP) though there are performance improvement methods, like the star scheme. In spite of the fact that MBD are still the most suitable for online analytical processing, now build in this opportunity relational DBMS or expand them (for example, MS Analysis Services or  Oracle OLAP Services are not the same that ROLAP).

Also there is hybrid online analytical processing of data (HOLAP) for hybrid products which can store multidimensional data naturally and also in relational representation. Access to MBD is provided using API for generation of multidimensional requests whereas to relational DB access is made by means of requests for SQL. An example of ROLAP server is Microstrategy7i Server.

The desktop OLAP tools (for example, BusinessObjects Explorer, Cognos PowerPlay, MS Data Analyzer) which is built in now EBIS facilitate to end users viewing and manipulation of multidimensional data which can arrive from server resource data of ROLAP or MOLAP. Some of these products have an opportunity to load cubes so they can independently work. As a part of EBIS these desktop tools are equipped with opportunities of server processing which go beyond their traditional opportunities, but do not compete with MOLAP tools. Desktop tools in comparison with MOLAP means have the small performance and analytical power. The interface through BusinessQuery for Excel, for example, of MS Excel2000/OLAP PTS, BusinessQuery for Excel is quite often provided. Practically all OLAP tools has Web expansions (Business Objects WebIntelligence for example), for some they are basic.

Corporate BI sets

EBIS is a natural way for providing BI tools which were delivered in the form of separate products earlier. These sets are integrated into tool kits of generation of requests, reports and OLAP. Corporate BI sets should have scalability and extend not only to internal users, but also on key customers, suppliers, etc. Products of BI sets should help administrators at implementation and management of BI without adding of new resources. Because of close relationship of Web and corporate BI sets some suppliers describe the BI sets as BI portals. These portal offers provide a subset of opportunities of EBIS using the Web browser, however suppliers constantly increase their functionality, bringing closer it to opportunities of tools for "fat" clients. Typical EBIS deliver Business Objects and Cognos.

BI platforms

BI platforms offer tool kits for creation, implementation, a support and maintenance of BI-applications. The appendices saturated with data with "custom" end user interfaces organized around specific business problems with the target analysis and models are had. BI platforms, though not so quickly grow and are widely used as EBIS, are an important segment thanks to the expected and already happening growth of BI-applications. Efforts of suppliers of the relational DBMS creating OLAP expansion of the DBMS many suppliers of platforms which provided multidimensional DBMS for OLAP to survive were forced to migrate to the area of BI-applications. DBMS product families providing possibilities of BI really push growth of the market of BI platforms. Partly it occurs thanks to bigger activity of a number of suppliers of DBMS.

Considering different tools, we see that EBIS are highly functional means, but they are not of such great importance as BI platforms or custom BI-applications. But BI platforms are usually not so functionally complete as corporate BI sets. When choosing BI platforms it is necessary to consider the following characteristics: modularity, a distributed architecture, support of the standards XML, OLE DB for OLAP, LDAP, CORBA, COM/DCOM and ensuring work in Web. They should provide also the functionality specific to business intelligence, namely: access to a DB (SQL), manipulation of multidimensional data, modeling functions, statistical analysis and business graphics. This category of products is represented by companies Microsoft, SAS Institute, Oracle, SAP and others.

BI-applications

BI tools are often embedded in applications of business intelligence (OLAP, generators of requests and reports, simulars, statistical analysis, visualization and data mining). Many BI-applications take data from ERP applications. BI-applications are usually focused on specific function of the organization or a task, such as analysis and sales forecast, financial budgeting, forecasting, risk analysis, analysis of trends, "churn analysis" in telecommunications, etc. They can be applied and is wider as in case of applications of performance management of the enterprise (enterprise perfomance management) or the balanced scorecard (balanced scorecard).

Investigation of data

Investigation of data (data mining) represents process of detection of correlation, trends, templates, communications and categories. It is executed by a careful research of the templates given using sensing technologies and also statistical and mathematical techniques. At investigation of data different transactions and conversions over crude data (selection of signs, stratification, a clustering, visualization and regression) which are intended are repeatedly executed:

1) for finding of representations which are intuitive for people who, in turn, understand the business processes which are the cornerstone of their activity better;

2) for finding of models which can predict result or value of certain situations, using historical or subjective data.

Unlike use of OLAP investigation of data in much smaller degree goes the user, instead it is necessary on specialized algorithms which set a ratio of information and help to recognize important (and earlier unknown) the trends free from bias and assumptions of the user.

Other methods and means of BI

Except the listed tools, the following means of the analysis can be a part of BI : packets of statistical analysis and time series analysis and risks assessments; simulars; packets for neural networks; means of a fuzzy logic and expert systems. It is in addition necessary to note means for graphic design of results: means of business and scientific and technical graphics; "panel boards", means of analytical cartography and topological cards; visualization tools of multidimensional data.

See Also

Business Intelligence determination

Standard blocks of the modern BI systems

Advantages of use of a BI system

Recommendations about the choice of BI

Implementation of BI. Typical errors

Business Intelligence, BI (world market)

Business Intelligence (Russian market)