Can In-Database Analytics contribute to Manufacturing QA?

Data Scientist, IBM Analytics, IBM

All too often, the morning news brings us a headline about another potentially catastrophic failure in a manufacturing part or process, such as the recent mid-flight airliner fuselage crack or tire failures associated with vehicle rollovers.  Certainly, manufacturers have many processes in place to detect and resolve defects, particularly those that might endanger the public.  While statistics and other analytical techniques are part of those processes to understand the root causes of defects, in many cases such analysis is done outside the operational environment and thus may take some time for findings to be translated into corrective actions.

By bringing the analytics into the operational environment, in-database analytics, particularly data mining, can extend and enhance the predictive and descriptive (“discovery”) power of analytics for manufacturing quality assurance (QA).  For example, data mining techniques such as clustering and associations can be used either interactively by analysts who need to work with the very latest available data or in automated analytical processes to detect and report emerging problem areas. 

We can expect to see a growing usage of in-database analytics, along with increasing automation of analytical and alerting processes, as data volumes grow and the need for faster data-to-results becomes ever more important.  In-database analytics are particularly suited to these needs in manufacturing QA as well as many other industries and areas of application.