[MUSIC] In this second lesson of the module, we will be exploring a high level understanding of advanced analysis, a key component of the digital manufacturing and design paradigm. Upon completion of this lesson you'll be able to state a generic definition of basis data analysis. Explain with at least one example, the need for advanced analysis techniques in the advanced manufacturing enterprise. Jeff Immelt, chairman and CEO, of General Electric has made this comment. If you went to bed last night as an industrial company, you're going to wake up today as a software and analytics company. The change is happening in front of us. This observation is a clear call to action to transform manufacturing to a new state, that uses the data generated in the digital manufacturing process, to more efficiently produce higher quality products. Before discussing an analysis we need to establish definitions. First is data. Data is unprocessed raw input. Some example of data are, in a class, each student's individual test score. A measured dimension that results from an inspection process like the diameter of a whole on a manufactured part. Information is data that has been processed to provide additional insight. Instead of a single student's grade, the average score of a class or of the entire school, is information that can be derived from the given data. In manufacturing, tracking how a dimension changes across multiple products, like if a diameter of a whole is consistently getting smaller, represents usable information. Data analysis is the act of transforming data into usable information. This information can be grouped together to establish long term findings that represent knowledge. This knowledge becomes the basis for taking action, like replacing or sharpening a cutting head if dimensions are tracking in a consistent direction that indicates wear. Data, by itself, is just a number and is not very useful. Data becomes powerful once it is processed and put into context. Like the saying goes, data rich, information poor. We can recall some simple data analysis methods from statistics. The mean is what typically comes to mind when you're asked to find the average of a set of data, and is found by adding up all of the data points and dividing by the number of data items. Median and mode, are two other approaches to developing a average for a data set, representing the middle value and the most commonly found value. While the range of a data set is the difference between the highest and lowest value. These statistics, along with slightly more advanced information, can measure how much variance or deviation is present in a data set. When coupled with knowledge about a manufacturing process, these simple pieces of information can guide a process improvement strategy like Six Sigma. Another example of how data and information drive action is using frequency, or how many times a particular value occurs in a data set, to focus on a high frequency occurrence first. With the success of Six Sigma and other process control strategies, a natural question may be, why would we need advanced analysis? The short answer is that the complexity of analysis required typically follows the complexity of the data or the product. So simple data analysis methods can be used for simple data analysis. But as we look to increasingly complex problems, like the supply chain for a complex product, we need more sophisticated approaches to generate actionable information. Technologies that have resulted in significant improvements in advanced analysis are, high performance computing and big data. High performance computing, also known as HPC, can enable advanced analysis. Sometimes, the analysis techniques required to process data, require computational power far beyond what our desktop computers are capable of. In such cases, we use high performance computing to analyze data, and convert data into information quickly, efficiently and with high reliability. Big data refers to the overwhelming amount of digital data, over 2.5 quintillion bytes of data every day that comes from everywhere. Digital content like pictures and videos, posts to social media, product design requirements, supply chain and purchase transaction records, as well as sensors in manufacturing equipment and products operating in the field. Big data spans three main dimensions, volume, velocity and variety. Volume refers to the constantly increasing amount of data. With companies accumulating terabytes of raw data and information as part of the design and manufacturing process. Velocity captures the concept of time sensitive data. Using production sensor data to predict that a piece of manufacturing equipment needs serviced can prevent scrap, rework or even down time that can result from an unexpected failure. Variety recognizes that there are many sources of data. Some, well structured and annotated, others that may require processing to become meaningful sources of information. An example of variety would be sensor data from production machines being coupled with video feeds from production lines. Advanced analysis offers opportunities to find insights and opportunities for improvement in manufacturing like, models could be developed for product designs and definitions. Rather that diagnosing a failure, prognostics or predictive analysis, could forecast when a piece of production equipment will go out of tolerance. Information about a jet engine could be used to predict it's performance, and when coupled with operational data, be used to generate a customized preventative maintenance schedule while in service. Advance analysis also enables design for x. Where the x stands for manufacturing, safety, reliability, performance, quality, etc. This lesson is intended as a basic introduction to the topic. To learn more, additional resources have been made available on the course website. Dr. Rowell Rye, an expert in this topic, walks through a more detailed discussion in course four of the Digital Manufacturing and Design Specialization Advanced Analysis.
