“If you know yourself and the enemy, don’t fear the result of a hundred battles. If you know yourself but not the enemy, for every conquest achieved you’ll also suffer a loss. If you know neither yourself nor the enemy, then you’ll be defeated every time.” This is one of the best-known quotes of The Art of War, by Sun Tzu. Of course, it was originally written for being successful in battles, but several businesspeople and coaches use this knowledge for building new commercial and logistical strategies.
In the world of supply chains, this knowledge of your organization, your production, your market, and competition, among others, is offered by statistics. It is almost impossible to know the organizational problems in your supply chain, to plan new production and marketing strategies, or to know what you should do to reduce industrial safety risks if you do not have the right information. You need to know yourself, know your battlefield and the problems you face. An excellent way to do this is by using data mining.
When we talk about data mining, we usually refer to methods of extracting information, and the systems derived from it, such as automatic learning. It is also common to relate the term “big data” to specify aspects of data mining, but there is still much confusion about the term (even though it was really introduced by Grigory Pyatetskim-Shapiro in 1989,) perhaps because of its novelty and, of course, the just recent implementations in various industries.
Data mining is a collective name used to denote a set of methods for detecting previously unknown, non-trivial, practically useful and accessible interpretations of knowledge necessary for making decisions in several spheres of human activity.
The basis of data mining techniques is all sorts of methods of classification, modeling, and forecasting, based on the use of decision trees, artificial neural networks, genetic algorithms, evolutionary programming, associative memory, and uncertain logic. Data mining methods are often referred to as statistical methods (descriptive analysis, correlation and regression analysis, variance analysis, factor analysis, component analysis, discriminant analysis, time series analysis, survival analysis, or link analysis.) Such methods, however, assume some a priori notions about the data being analyzed, which is somewhat inconsistent with the goals of data mining (the discovery of previously unknown non-trivial and practically useful knowledge.)
Read also: Big Data: The great revolution of supply chain statistics, by David Kiger
Data mining has evolved a lot in recent years thanks to the technological advances, and, today, it is a very extensive area. Proof of this are the job vacancies, the most demanded profiles of the industry, and, of course, the increase of possibilities to formative level which are appearing to which must be added the proliferation of conferences that take place annually around the world, both focused on the methodology from a more technical point of view and business oriented, in which the topic of data mining is usually dealt with in conjunction with metrics and analytics.
The field of data mining from a statistical point of view and applied to the logistics of the supply chains can be researched thanks to the algorithms or the methods applied. Data mining information can help to optimize processes in any of the following areas:
- Quality control.
- The level of errors and their frequencies.
- The times for changing tools.
- The productivity levels of different processes, activities, and products.
- The levels of satisfaction of customers and users.
- Types of accidents and their frequencies.
- Pareto analysis of defects, costs, profitability, and sales.
- Sales by customers, vendors, zones, and products.
- Predictions of sales by zones, products, services, or branches.
- The capacity of processes in terms of costs generation, quality, and productivity levels.
- Total times of productive cycles.
- Response times.
- Inventory management.
Suppose you run a chain of restaurants (an example that includes complex logistics, indeed.) Suppose the company grows some of the food. How can you tell if a pest is affecting the crops? How can you know which is the most ordered dish for your customers in the moments of most traffic? Even how can you find out how potential customers get to your business through Internet searches? All this, all the details that make up the supply chain, and more, can be supplied through the implementation of Big Data and data mining.
A typical data mining process consists of a series of general steps. First off, the largest amount of data is collected using Big Data technology, then analyzed (in graphs, Excel tables, and other metrics,) then the data is interpreted (the patterns, sequences, and other behaviors of the processes,) and this is useful to produce the necessary knowledge to improve the logistics of the organization (see more in this video.)
In order to run a business, it is much better to have a little light to know the problems instead of being guided by common sense and intuition. Every day, you need to be more productive and to systematically eliminate waste. For making it possible, you require accurate information. Stop running your business like it was fifty years ago!
Recommended: Data Mining Trends for 2017
* Featured Image courtesy of JD Hancock at Flickr.com