AI-Enabled Analytics for Business. Lawrence S. Maisel. Читать онлайн. Newlib. NEWLIB.NET

Автор: Lawrence S. Maisel
Издательство: John Wiley & Sons Limited
Серия:
Жанр произведения: Зарубежная деловая литература
Год издания: 0
isbn: 9781119736097
Скачать книгу
as depicted in Figure 1.1, as well as other forms of AI (that is, any other technology/technique that enables a machine to make a human decision).3

Schematic illustration of superset and subsets of AI.

Schematic illustration of ML process.

      Another set of terms to get our arms around is analysis and analytics. Analysis, in business reporting, involves calculations of arithmetic (add, subtract, multiply, and divide), whereas analytics for business encompasses mathematics (algebra, trigonometry, geometry, calculus, etc.) and statistics (about the study of outcomes).

      In a profit and loss statement, there is a variance analysis of current year actual performance against budget. The analysis is expressed as the difference in dollars and as a percent. The variance analysis uses arithmetic to make a measurement of the existing condition of the company compared to what it planned for the year. This analysis is comparative information from arithmetic on data and descriptive of a current situation, but it is not an insight that is additive to a decision.

      Insight, as defined with respect to the value from data, is that not known about the business and when known should affect decisions, and insights are derived from analytics that applies mathematics to data.

      For example, say sales are down 15% for the past three months, but sales are predicted to increase this month. This prediction is based on a correlation of unemployment as a three-month inverse leading indicator to sales, meaning as unemployment goes down, sales will go up. In this example, unemployment has been dropping for the past three months, so the prediction is for sales to increase in the current month.

      The use of correlations to make a prediction is analytics that reveals an insight, which was not known from the data or information from the analysis of the data, and which when known will affect decisions. In this case, without knowing the prediction of the lead indicator, the business would run deep discounts to attract sales. However, knowing that sales are predicted to reverse direction would cause the business not to discount or to only offer small discounts.

      Accordingly, analytics can powerfully reveal unbiased insights, as it applies mathematics on data that is void of the personal and political pressures that are exerted on humans when they make forecasts and predictions. As humans, we want the future to be what we desire or what we need, so we can make any forecast come to our desired outcome. As such, analytics is especially potent to enable unbiased data-driven decisions.

      Business intelligence (BI) tools date back to the 1980s and enabled multidimensional reporting. BI went beyond spreadsheets to ingest large amounts of data from several data sources and then segment (into separate dimensions) the data into hierarchies. This approach gave users the ability to organize and dive into more data more intelligently.

      Today, legacy BI tools have essentially become data-marts for data extraction into spreadsheets for reporting. BI tools are largely maintained by IT and require programming to build cubes (specialized BI databases) to respond to predefined questions. However, legacy BI is too rigid and complex for most users, so IT departments often program user-requested reports and data extractions (for download to other applications).

      The complexity of BI gave birth to data visualization tools that were introduced in the 2000s and offered graphic representations of data in many forms, often combined into dashboards to render a story about key aspects of the business. Dashboards can be informative but typically not analytical.

      Dashboards are of prime value to combine visual charts with tabular data of KPIs and key values for comparisons.

      The picture below is … where data and images of trends can work together to offer a view to the past and present. Like a car's dashboard, the numerical readings at the top tell key performance data needed to be known; e.g. if we're running low on gas…

A photograh of a tablet screen shows a few graphs.