Business Intelligence becomes more and more an essential technique for a company which wants to use its information in the best way possible. Within this topic the idea of „Predictive Analytics“ is to forecast complex economic correlations and use this knowledge to make better decisions within a company. This extension gives the opportunity not only to answer relevant questions within a company, but to predict future happenings to prevent unpredictable situations.
Callbacks within the automobile industry, winter chaos for public transportations or the financial crises; customers wondered: why do companies not recognize these developments through internal analyzes way ahead? Business Intelligence methods can answer questions concerning the company’s current situation quite precisely. The combination of target figures and actuals give a solid foundation to support management decisions. Even if the systems are working with reliable and right data it allows an insight into future developments. But if market places and business dealings are not rotating in the same direction it needs specific analytic tools to receive unerring predictions. Business Intelligence is mainly concerned with the events in the past, and their effects on the present. Predictive Analytics on the other hand deal with happenings in the future which shall be prevented through data analysis. This is why “Predictive Analytics” play a decisive role and expense the business intelligence method by giving an insight into the future and becomes one of the most important Big Data trends.
Predictive Analytics starts where OLAP and Reporting ends: instead of analyzing existing situations Predictive Analytics is trying to make predictions of possible happenings in the near future based on data models. Thereby a close connection to Data Mining is noticeable. But are Predictive Analytics and Data Mining the same? In most cases the two terms are used as synonyms and indeed Data Mining plays an important role within Predictive Analytics solutions, but Predictive Analytics transcends Data Mining and uses further methods. It uses for example text mining (algorithm based analysis method) to find structures within non structured text parts such as articles, blogs, tweets or facebook postings.
To use the Predictive Analytics in a most efficient way it is necessary to include the existing database of the data warehouse. Most companies also use synergistic effects by integrating the functions of predictive analytics into an existing Business Intelligence environment.
The Analytics maturity model (Analytics-Reifegradmodell) by Gartner gives an insight into the existing analytics methods and offers the opportunity to rank Predictive Analytics. There are four different methods: Descriptive Analytics (“What happened?”), Diagnostic Analytics (“Why did it happen?”), Predictive Analytics (“What will happen?”) and Prescriptive Analytics (“How can we make it happen?”).
Predictive Analytics is not only used for security (Predictive Policing) anymore but in many other business lines as well. For example within the energy industry: the smart grid of the future provides load forecasts and predicts the demand for electricity to harmonize the power consumption and the fluctuating power generation (helioelectric power plant and windmill). Another example are the banking institutions: they estimate the credit scoring from the probability or the risk with which a customer could not afford the future installments of a loan granted. Basically it can be said that Predictive Analytics is an ongoing and iterative process because with every adjustment forecasts are getting more and more precise.
Koeffer, Sebastian (2014): Mit Predictive Analytics in die Zukunft blicken. http://www.computerwoche.de/a/mit-predictive-analytics-in-die-zukunft-blicken,2370894 (14.04.2016).
Mauerer, Jürgen (2015): Was ist was bei Predictive Analytics?. http://www.computerwoche.de/a/was-ist-was-bei-predictive-analytics,3098583 (14.04.2016)