More and more devices, manufacturing tools and factories are equipped with sensors that can collect huge amounts of data about themselves and their environment and connect to the data science cloud. This is called Big Data, but why is it necessary to implement data science in an industry? Implementing Big Data in industry turns data into knowledge and a clear competitive advantage. Experts believe that Big Data systems, in addition to managing and processing large amounts of data in a computationally efficient manner (speed), "allow for the incorporation of other data analysis tools, in addition to the fact that any open source enhancements are incorporated progressively into information management or analysis." In the same vein, some professors of artificial intelligence believe that "raw data adds almost nothing". Thus, they consider that it is thanks to the consolidated techniques of data analysis and the technologies that support them, that we can make sense of them and obtain conclusions and advanced knowledge capable of helping the industries that integrate them. It is necessary to promote the deployment of data science in companies and industries, because without such solutions, it is practically impossible to take advantage of the data collected. This also implies a cost that makes no sense if it does not bring anything.


It is precisely those industries that manage to collect their data and analyze it correctly that will get what some call "the oil of the 21st century". Indeed, the industrialization of data, will offer advanced knowledge and clear added value while being able to use this data to learn from it and detect opportunities, hidden relationships and perform knowledge management. At the same time, there is a fundamental advantage for these companies, as they can perform both diagnosis and corrective action when problems are detected. A competitive advantage cannot be achieved if data is simply stored, but not analyzed to try to detect possible deviations or anomalies. To be able to anticipate many problems, one must be aware of the warning signs that may occur and, to do this, one must know how to find these signals among the huge amounts of data available.


But what about those who don't? Where are they and where are they going? Indeed, most of them are on the "dark side". Thus, they rely on traditional models and use biased knowledge, based on the knowledge of specific people and without knowing whether it is true or persuasive. As a result, they do not have a defined model and therefore do not know which parts of the process generate uncertainty, how one variable or another affects quality, or how a raw material is likely to generate one defect or another when there are changes in the process; when a system may fail early and why. They typically waste resources either by deploying data capture and storage systems that don't prove useful or by not having them and "blindly" trying to solve certain problems without data and/or evidence guiding the solutions.


Today, the demand for training in data analysis on the Big Data paradigm is high, and it is a skill that is highly valued by industrial companies, which implies extending training to managers of non-IT functions, to people who know the processes, what to look for and how to do it. In addition, other profiles more adapted to architectures, can integrate new data science devops skills in their organizations and know the analysis tools, which can then be used by other teams in your company. Multidisciplinary teams, bringing together different profiles and knowledge around the Big Data paradigm, will allow companies to be masters of their knowledge and to make decisions, not based on experience, but on the reality of their business and business demand.