Business Intelligence (BI) is a broad term but is generally accepted to refer to the technologies and techniques used to capture, store and analyse data. The goal of BI is to assist organisations in making decisions. Business Intelligence and Business Analytics are terms that are often used interchangeably but Business Analytics can be thought of as being more applied in nature. To distinguish between the two, BI can be defined as the querying, reporting and presentation of historical data while Business Analytics can be defined as the applied data, statistical and quantitative techniques used to deepen insights into organisational performance and assist in business planning. BI itself is sometimes thought of as simply reporting but provides a much richer suite of tools and techniques than just reports. BI is used to analyse and uncover information about past performance using data stored in a database, or more frequently, a data warehouse.

The most well known technique in Business Analytics is Data Mining. Data Mining is the process of sifting through data to pinpoint the patterns in same. Using Data Mining techniques, data can be classified or grouped and statistical predictions can be made. Data Mining and statistical techniques are often considered to be one and the same but statistics is only one discipline that is used when Data Mining. Data Mining is in fact as a blend of statistics, Artificial Intelligence (AI), Machine Learning and Database Technologies. There are many techniques used in data mining. We will consider the most popular techniques here.

Artificial Neural Networks are modelling techniques based on the learning process of the human brain. Using Neural Networks, learning from existing data can take place and predictions can be derived from this data. Genetic Algorithms are optimization techniques that are based on the concepts of natural evolution. Essentially, it is a search technique that is used to find exact or approximate solutions to optimization and search problems. Decision Trees are tree-shaped structures that represent sets of decisions and provide a set of rules that can be used to predict which data records will have a given outcome.  The Nearest Neighbour Method is a machine learning technique that uses the ratio of the expected and observed mean value of the nearest neighbour distances to determine if a data set is clustered. A statistical test of significance of the near neighbour statistic is used to quantify the departure of the pattern from random. Another statistical technique is rule induction which is used to extract rules from data based on statistical significance. Finally, visualisation techniques use graphical software to illustrate the relationships between data.

Data Mining, then,  is clearly more than statistical analysis of data. It is the process of analysing data from different perspectives and transforming it into useful information that can be used to improve business decision making, reduce costs and increase revenues.  Other techniques used in Business Analytics include simulation, forecasting, optimization and experimental design. Most of these techniques are rooted in the field of Management Science, better known as Operations Research. 

As noted, Business Analytics is a distinct field from Business Intelligence. The latter is focused on extracting data on historical performance. One common technique used in Business Intelligence is OLAP (Online Analytical Processing). OLAP is a data structure that is commonly referred to as a cube. Using a data cube, data can be viewed in multiple dimensions e.g. sales in a branch of a store for a particular year is data that can be considered multi-dimensional. Data cubes are usually constructed within Data Warehouses and Data Marts and for this reason BI and Data Warehousing are often viewed as complementary technologies. However, it is not necessary to have a Data Warehouse as the underlying data storage for BI.

Another key technique for BI is querying. Querying is the extraction of information for business decision making from a database or data warehouse. There are many BI querying tools in the marketplace but key determinants of the effectiveness of such tools are that they be database independent, can support an industrial strength database that may consist of millions of rows of data and hundreds of columns, be easy to use and be easy to integrate with a spreadsheet or Enterprise System. Reporting is often considered to be synonymous with BI but it is actually just one facet of the area. Reports can be text-based or graphical and are key tools for Performance Management. In fact, the areas of BI and Corporate Performance Management (CPM) are gradually converging and the tools and techniques used for both disciplines are becoming inextricably linked. IBM Cognos (http://www.iba-it-group.com/en/services/BI/) is just one example of a tool that combines the two areas. 

The final area of Business Intelligence we will consider is that of Alert and Exception Reporting. This area makes BI more than just a measurement tool for past performance. BI is also effectively a Business Process Monitoring system which highlights an exceptional event occurring in business operations in near real-time. Rules can be implemented to set up criteria for exceptional events. These events can be managed by the BI system to ensure that appropriate responses are carried out and users can be notified of the alerts through web portal headlines and/or their email system.

As noted, BI and Data Warehousing are technologies that are frequently linked. Extract, Transform and Load (ETL) is sometimes considered as a BI technique but more often relates to Data Warehousing. In essence, ETL is the extraction of data from different sources and systems, the cleansing and reformatting of this data and the loading of this data into another database, data warehouse, data mart or Information System.

There are many vendors in the Business Analytics and Business Intelligence spaces. SAS (http://www.sas.com/), Business Objects (http://www.sap.com/solutions/sapbusinessobjects/index.epx) and Crystal Reports (http://www.crystalreports.com/)  are all leading Business Intelligence and/or Business Analytics software. Computer Programmers can also integrate BI tools into their own solutions though the use of tools such as the Eclipse BIRT project for J2EE (see http://www.eclipse.org/birt/phoenix/).

The question we must how consider is how Business Analytics and Business Intelligence are of use to smart objects and smart networks.  Smart Objects (RFID, Smart Meters, Wireless Sensor Networks and GPS among others) provide a new pool of data for organisations. Given the potential volumes of devices in smart ecosystems it is clear that analysis tools and techniques are required to query and report on the data produced by these networks. In the case of smart meters and the smart grid, the raw data ultimately determines revenues so BI can be used to assess revenue performance and/or cost reductions for non-utilities. A central system for reports, Key Performance Indicator (KPI) measurement and data analysis simplifies the collaboration and sharing of information from smart ecosystems while the Business Analytics techniques we discussed can provide insight and patterns into the vast volumes of real-time data produced and also enable prediction models to be built. 

Smart Objects and smart networks should not exist in isolation in an organisation. The data these devices and networks produce needs to be aligned with the overall goals of the organisation. Business Intelligence and Business Analytics are key tools for the measurement of performance, the setting of objectives and the ability to make business decisions in a timely fashion. The pools of data provided by smart objects can be thus transformed into meaningful information for assessing past performance and guiding future planning and decision making. The data from different smart networks can also be combined by BI and Business Analytics tools to give a better picture of what is occurring within the ecosystem.

In summary, BI and Business Analytics transform the data from smart objects into meaningful information for business and makes this information available to all the key decision makers within the organisation. However, BI tools do not generally have the capacity to capture the data from smart objects themselves. Vertoda Middleware would be required to perform this function. Data is captured and stored as meaningful information in a database or data warehouse. The information produced by Vertoda can then be accessed by any type of BI or Business Analytics system and the benefits described above can be derived.