Objectives and architecture of BI systems

2.2.1 Objectives of BI systems

BI systems may be analysed from different perspectives. Decision makers and organisations should predominantly associate BI with organisational implementation of specific philosophy and methodology that would refer to working with information and knowledge, open communication, knowledge sharing along with the holistic and analytic approach to business processes in organisations. According to Olszak and Ziemba (2007), BI systems are considered to be solutions that are responsible for transcription of data into information and knowledge and they also create some environment for effective decision making, strategic thinking and acting in organisations, as it is depicted on the Figure below.

Fig. 2. The role of BI systems in decision making Source: Olszak, & Ziemba, 2007

Observation of different cases of BI Systems allows for stating that the systems in question may support data analyses and decision making in different areas of organisation performance, particularly including the following (Hsu, 2004; Olszak, & Ziemba, 2003):

• Analysis of the company's financial activities, including cost and income analysis, calculation and comparative analysis of the company's income statements, analysis of the corporate balance sheet and profitability or loss, analysis of financial markets, risks analysis;

• A variety of marketing analyses that includes analysis of all sales revenues, sales profitability, profit, achieve sales targets, time of orders, actions of competitors, stock quotes changes;

• Analysis of customers and suppliers, which concerns the time and quality of contact with customers and suppliers, profitability from customers, modeling of customer behavior and reactions, customer satisfaction, etc.;

• Analysis of production control that allows you to identify production bottlenecks and pending orders, which allows organizations to examine the dynamics of production and to compare production results obtained by departments or plants, etc.;

• Logistics analysis to quickly identify the best supply chains;

• Analysis associated with wage data, including analysis of performance of employees, wages and salaries in breakdown by kinds of employment, payroll surcharges, personal contribution reports, analyse of average wages, etc.;

• Analysis of personal data, which include the study of employment turnover, types of employment, provision of information on personal data of employees, data security, etc.

2.2.2 Core components of BI architecture

BI requires analysts to deal with both structured and semi-structured data (Rudin and Cressy, 2003;

Moss, 2003). The term semi-structured data is used for all data that does not fit neatly into relational or flat files, which is called structured data. While Data Warehouses (DW), Enterprise Resource Planning (ERP), Customer Relationship Management (CRM), and databases mostly deal with structured data from data bases, the voluminous semi-structured data within organizations is left behind (Solomon Negash, 2004). To create business intelligence information, the integrated data are searched, analyzed, and delivered to the decision maker. In the case of structured data, analysts use Enterprise Resource Planning (ERP) systems, extract-transform-load (ETL) tools, data warehouses (DW), data-mining tools, and on-line analytical processing tools (OLAP).

A typical architecture for supporting BI within an enterprise is shown in Figure 3 (Chaudhuri, S., Dayal, U., & Narasayya, V., 2011). The data over which BI tasks are performed often comes from different sources — typically from multiple operational databases across departments within the organization, as well as external vendors.

Fig. 3. Typical business intelligence architecture Source: Chaudhuri, S., Dayal, U., & Narasayya, V., 2011 Data storage

The data over which BI tasks are performed is typically loaded into a repository called the data warehouse that is managed by one or more data warehouse servers. Data should reflect the current, real and complete picture of the business. Information in the data warehouse (including historical data) is collected from various operating (transactional) systems and structured in a special way for more efficient analysis and processing of requests (in contrast to conventional databases, where the

information is organized in such a way as to optimize the processing time of current transactions).

Data warehouses contain huge amounts of information covering all available aspects of the enterprise and allowing to consider all aspects of the business in the aggregate. Subsets of data – the so-called data marts (data marts) - can be extracted from the common storage to solve narrower, specific tasks. The specific data needed for BI are downloaded to a data mart used by planners and executives. Outputs are acquired from routine push of data from the data mart and from response to inquiries from Web users and OLAP analysts (Solomon Negash, 2004).

Data integration

To form and maintain data warehouses, the so-called ETL tools are used – tools for data extraction (extract), data transformation (transform), that is, bringing them to the required format, processing in accordance with certain rules, combining with other data, etc., as well as for loading data (load), writing data to the storage or to another database. In addition to ETL, BI systems include SQL tools (structured query language) that allow users to access data directly.

To integrate data from disparate sources, modern BI systems use an intermediate, virtual metadata layer, which eliminates the need for business users to understand the intricacies of storing and processing information and facilitates changes. These tools do not require any physical operations to move and process data, which distinguishes them from ETL tools. The use of such a metadata layer, in principle, eliminates the need to organize expensive data stores (but it is necessary to take into account the issues of ensuring the necessary performance). This approach to integration of Analytics from TEC (Technology Evaluation Centers) is defined as EII (Enterprise Information Integration).

In addition, enterprise portals can be created for data integration to provide connectivity at the data and business process levels. Such portals implement only the external relationship, in other words-provide shared access to information. This implementation of the experts from the TEC was called EAI (Enterprise Application Integration).

Real-time data warehousing often relies on very fast ETL-style updating. Real-time analytics focuses on constant event monitoring and predicting as well as prescribing specific actions to be taken in order to take advantage of an opportunity or to mitigate a problem (Power, D. J., & Sharda, R.,2015).

Data Analysis

For a comprehensive analysis of data in modern BI used OLAP tools. They allow you to consider different data slices, including time, allowing you to identify different trends and dependencies (by region, product, customer, etc.). Various graphical tools are used to represent the data – reports, graphs, charts, customizable with the help of various parameters.

Blocks for deep data exploration (data mining) are included in the most advanced BI solution.

Sometimes this term is mistakenly used to refer to tools that allow to present (display) information in a new way, but in fact these tools are designed to help in identifying hidden (non-obvious) patterns, models, forecasting. They are based on the scanning and statistical processing of huge amounts of data and are ultimately designed to facilitate the adoption of correct and informed strategic decisions through the analysis of various scenarios. Neural networks and decision trees are used as tools.

Data Visualization

Common means of data visualization in modern BI-solutions are information (control, instrument) panels (dashboards), where the results are displayed in the form of scales and indicators that allow you to monitor the current values of the selected indicators, compare them with the critical (minimum\maximum) values and thus identify potential threats to the business. Operational BI output often focuses on a dash board of performance metrics. Strategic BI is more likely to present managers a scorecard of key performances indicators with historical and goal comparisons (Power, D. J., & Sharda, R., 2015).

Control panels are considered one of the most convenient ways to present information about the

"state of health" of the business. They allow you to fit on the screen all the important information about current operations, identified and potential problems. Control panels, as well as scorecards, are based on the analysis of key performance indicators (KPIs). However, as a rule, control panels show the current state of the General indicators, and indicator maps are designed to compare the current indicators with the planned, target, and display the dynamics of changes in these indicators over time. Scorecards are usually more personalized, customized depending on the roles and tasks of a particular user (financial management, procurement, sales, etc.). If necessary, all these indicators can be detailed with the help of additional reports, graphs and charts.