“Data mining” can be described as the process of finding patterns and relationships in data. There are several different flavors of data mining, using different methods and pursuing different results. For example, a data mining project might look for:
- associations (interconnected events or information items)
- sequential relationships (one event leading to another)
- affinities (events that occur in clusters, information items that are frequently found together)
There are also two levels of approach to data mining. “True” data mining is an IT-intensive process, using complex algorithms and very sophisticated procedures to discover deep and/or unexpected patterns in data. At the business level, however, data mining is often viewed more generally, as a way of exploring data to answer questions and support analysis.
The term “desktop data mining” can refer to utilizing the result of data mining projects to develop business intelligence, by means of a desktop tool or end-user interface. In this sense, large data mining operations can take place in the database or data warehouse, and the information extracted in that process can be made available to business users for further analysis. But the term may also be used casually to refer to any kind of data exploration performed by business users–such as testing what-if scenarios, looking for common factors in customer behavior, or probing for trends in purchase data.
Why does it matter?
Originally, data mining techniques were utilized mainly for scientific and mathematical research, where it was necessary to work with very large amounts of data and/or many variables. In recent years, though, data mining has become an indispensable tool for business. The application of data mining concepts to customer information has already revolutionized CRM and marketing–and as businesses accumulate more and more data, the use of data mining has expanded across the enterprise.
For example, the insights gained through data mining might lead to changes in the way a company captures and organizes information, thereby optimizing data processing and producing better, faster answers to business questions. Data mining can also be applied to operational and financial data, helping to identify economies and efficiencies. And since precise, reliable forecasting is so important throughout the enterprise today, predictive analytics has become an especially strong area for the development of data mining applications.
Data mining is also increasingly important because of its potential applicability to the very large amounts of unstructured data derived from social media and the web.
The availability of data mining results and/or tools has empowered business users to perform more—and more powerful—analytics right at their desks. With user-friendly interfaces, strong data visualization capabilities, and built-in collaboration options, today’s analytical platforms are transforming the way information drives business.
Development trends in data mining include the expansion of “mining” techniques to various types of unstructured data (for example, text mining, web mining, sentiment mining, etc.), as well as the integration of data mining with artificial intelligence. For a nice outline of some front-burner issues in these areas, check out the Topics of Interest list at Social Media, Data Mining & Machine Learning. There is also a gathering momentum for the development of “spatial mining,” which integrates data mining with GPS data and maps.
For an extensive (and free!) introduction to data mining concepts and applications, download chapters from Data Mining: Know It All at SearchBusinessAnalytics.com.
Subscribe to our blog to stay informed on desktop data mining and other data analytics topics.