Planning Data Configuration for Statistical Analysis

Article

Sarell, William E.   (2000, ASQ)  

Quality Progress    Vol. 33    No. 7
QICID: 13934    July 2000    pp. 39-45
List $10.00
Member $5.00

FOR A LIMITED TIME, ACCESS TO THIS CONTENT IS FREE!
You will need to be signed in.
New to ASQ? Register here.

Article Abstract

A greater availability of electronically collected data has allowed problems to be solved on the basis of data rather than intuition. The combination of quality initiatives with computerized data analysis packages means increased utilization of statistics to analyze process data. Because historical data tends to be collected on a moment-to-moment basis, however, the data is not independent, and therefore, not appropriate for basic statistical analysis. Practitioners must determine how to arrange data so they become independent, allowing the effective use of statistical analysis tools, such as control charts, histograms, and scatter plots. To determine the best method for getting data into a form that can be analyzed by such tools, it is useful to visualize the process first. After visualization, the process can be divided into logical segments, or phases, and process parameters can be chosen for acquisition. After planning, data interfaces can be developed to move the data into a statistical database for analysis. The use of statistical tools began in the manufacturing industry where the critical measurements of each part manufactured are independent of each other. This type of data can easily be analyzed using various statistical tools. In contrast, critical measurements and variables in the bulk chemical manufacturing industry are measured continuously, with successive measurements based on those that occurred previously. Chemical batch processes are complex and cannot simply be summarized. The critical parameters of a process can be analyzed statistically, reducing process variation, if data are collected from the process in a way that allows the process to be visualized as resembling a discrete manufacturing process. This is accomplished by viewing the process as a logical succession of events, or as a timeline in which events are arranged in an expected time order. After visualization, process parameters must be selected on the basis of the process expert's knowledge. A strategy for collecting data can be established when phases and parameters are clearly defined. A strategy for summarizing the data is also necessary. Using the skills of computer staff, a method for moving summarized data values into the statistical database can be devised. Guidelines for planning effective data collection and analysis involve the creating of a small group representing a cross-section of technical disciplines, collecting available flow diagrams of the process, compiling a list of all variables currently measured, developing timelines to show a logical view of the process, subdividing the timelines into functional process phases, determining parameters for each phase, documenting phases and parameters into a hierarchical table, and establishing strategies for moving parameter data from source systems to a statistical database.

Keywords

Chemical and process industries,Data analysis,Manufacturing,Statistical quality control (SQC),Process management,Statistical design,Process analysis


Browse QIC Articles Chronologically:     Previous Article     Next Article

New Search

Featured advertisers





ASQ is a global community of people passionate about quality, who use the tools, their ideas and expertise to make our world work better. ASQ: The Global Voice of Quality.