Due to the complexities of making products, most manufacturers are used to having large influxes of data from machines, processes, shipping, etc. What may be new to these companies, though, is having tools to retrieve actionable information from these piles of Big Data.
LNS Research and Mesa International teamed up to compile a survey of manufacturers on how they are using new technologies. Among the information gathered was how these companies felt they could use Big Data from the manufacturing plants and the overall enterprise. Of the more than 200 responses, 46 percent felt that Big Data analysis could help them “better forecast products” and production. Another 39 percent believed that Big Data mining will allow them to “service and support customers faster.” Other metrics from the survey include:
With companies looking to become more customer-driven, the responses are not really surprising. According to Manufacturing Business Technology Magazine, the ways that manufacturers could use Big Data production analytics for forecasting are numerous:
Here Big Data could operate in a myriad of ways, including identifying correlations between customer data, scheduling, and maintenance, which would have the potential to identify hidden patterns that could enable greater operational efficiency, better anticipate order lead times, shorten asset/machine downtimes, and make materials purchasing and WIP decisions more effectively.
The report also showed that only 5 to 6 percent of respondents felt that Big Data had no “future use or impact on their manufacturing performance.” That’s a good sign that the majority of companies are seeing Big Data technologies as an integral part of future manufacturing and corporate growth.
One example of a corporation that is already making use of Big Data to improve production is pharmaceutical company Merck. InformationWeek recently reported that the company set up Hadoop to analyze Big Data from 16 sources to investigate “higher-than-usual discard rates on certain vaccines” being manufactured, as well as reasons behind lagging yield rates on other vaccines.
Previous attempts at using spreadsheets to compare data and find the potential culprits took months and allowed researchers to look at only two batches at a time. When Merck Director of Manufacturing Advanced Analytics Jerry Megaro teamed up with VP of Information Technology for Merck & Co. George Llado, the two developed the process of using Hortonworks Hadoop distribution on AWS, and that was key to solving the mysteries of manufacturing the vaccines:
Megaro's team was then able to come up with conclusive answers about production yield variance within just three months. In the first month, July 2013, the team loaded the data onto a partition of the cloud-based platform, and it used MapReduce, Hive, and advanced dynamic time-warping techniques to aggregate and align the data sets around common metadata dimensions such as batch IDs, plant equipment IDs, and time stamps… Through 15 billion calculations and more than 5.5 million batch-to-batch comparisons, Merck discovered that certain characteristics in the fermentation phase of vaccine production were closely tied to yield in a final purification step.
Though their results may not make sense to those of us who aren’t in vaccine manufacturing, the takeaway is that Merck can apply the same technologies to other vaccines and by implementing what it learned, the company feels it can now advance the manufacturing of other in-development vaccines. So, Big Data analysis is helping Merck produce life-saving vaccines efficiently and in “a more plentiful supply.”