How Some Companies are Using Big Data in Manufacturing

Large-capacity mainframes, software that handles inconsistent information sources and improved abilities to create accurate computer models, fuel the ability of big data analysis to provide manufacturers with profitable means of streamlining production, improving quality, reducing costs and solving unanticipated problems. These improvements allow companies to improve profit margins even in difficult economic climates.

Companies such as Raytheon depend on consistency across all processes in the manufacturing of missiles. A single bad fastener can cause a multi-million dollar machine to fail. An industrial application of big data analysis compiles massive amounts of information on every tiny detail of the assembly process and yields software that will not allow the line to move forward unless employees attend accurately to the fine details, such as using the right hardware or even the right amount of torque on a wrench.

Bottlenecks in manufacturing increase the time needed for each final product, and big data analytic techniques can identify in detail exactly where these slow-downs occur through collecting information on the time needed for each step on a line. Harley Davidson reconfigured a line to ensure that motorcycle fenders were streamlined logically into the assembly process instead of being transported from one place to another on a cart. Sherwin Williams determined that machinery needed repair after 10,000 iterations. They began repairing at 9000 and reduced the amount of downtime for its machinery. Not only can analysis of data around bottlenecks yield information about suspected problems, but it can also identify unanticipated timing problems, the solution of which increases efficiency even more.

Increased computational power allows industrial leaders to examine multiple variables that may have an influence on the quality of a final product. Pharmaceutical companies need to produce consistent products on which peoples' lives depend, yet as many as fifty disparate factors contribute to the outcome of a particular lot. A combination of crunching fine-grained data and creating predictive models helps to narrow fifty variables to the four or five most significant ones, allowing companies to implement changes likely to make the greatest improvement in the resulting goods. The choice of which raw materials to use introduces another set of variables that can be examined through big data analysis and reduced through collecting large amounts of information on product processes and outcomes.

A final industrial application involves the use of existing historical data sets, new sources of data such as comments made about products on social networks, information on how employees do their jobs, technical specs on raw materials, and so forth, in order to identify unforeseen factors affecting quality, cost, and time. Modern software allows for the collection and investigation of information from a wide range of sources, which in the past could not have been combined. For example, data consumers may identify manufacturing defects that might not be detected by quality control procedures, or they may use big data analysis to respond to consumer concerns.

In challenging economic times, manufacturers must reduce the cost and the amount of time it takes to create goods. Tech savvy consumers and purchasers demand high-quality products and use customer reviews and other forms of communication to denigrate commodities that do not meet these standards. Big data analysis helps create efficient industries that meet buyers' requirements for quality.

(Photo courtesy of stockimages / freedigitalphotos.net)