MAKING SENSE OF DATA

Transcription

1 MAKING SENSE OF DATA Donald J. Wheeler SPC Press Knoxville, Tennessee

2 Contents About the Author Introduction ix xi Chapter One Why Continual Improvement? What is Quality? The Three Questions Summary 6 Chapter Two Visualize Your Process Brainstorming Flowcharts Constructing Flowcharts Deployment Flowcharts PERT Diagrams Do People Really Do This Stuff? Processes and Systems Cause-and-Effect Diagrams Exercises 27 Chapter Three Collecting Good Data What Are You Trying to Do? Operational Definitions Counts, Categories, and Types of Data 35 Chapter Four Visualize Your Data Running Records Bar Charts Pareto Charts Histograms and Stem-and-Leaf Plots Summary of Bar Charts and Running Records Pie Charts Exercises 63 Chapter Five Graphical Purgatory Computer-Generated Bar Charts Computer-Generated Running Records Stacked Bars and Linear Pie Charts Radar Plots Improving the Two Best Computer Graphs of Table Summary 80 v

3 Chapter Six Some Arithmetic Measures of Location Measures of Dispersion Numerical Summaries and Data Types Ratios, Percentages, and Proportions How Many Digits Do You Need? Comparing Apples and Oranges 91 Chapter Seven Visualize Your Process Behavior Two Types of Variation Two Types of Action The Basic Process Behavior Chart: the XmR Chart An XmR Chart for Accounts Receivable Exercises 104 Chapter Eight Interpreting the Process Behavior Chart What Do Predictable Processes Look Like? Detecting Strong Signals and Weaker Signals Why Johnnie Can t Breathe The Voice of the Customer and the Voice of the Process Summary Exercises 118 Chapter Nine Using XmR Charts Effectively Comparing Groups Using Process Behavior Charts The Has a Change Occurred? Chart Why Susie Can Breathe A Clinical Report: Peter s Story How Do You Use These Charts for Continual Improvement? 140 Chapter Ten How to Create Good Charts What Kind of Data Belongs on an XmR Chart? Using the Median Moving Range Charting Adjusted Data Exercises 151 Chapter Eleven What Makes the XmR Chart Work? The Logic Behind Process Behavior Charts Why Three-Sigma Limits? The Wrong Way to Compute Limits But How Can We Get Good Limits From Bad Data? So Which Way Should You Compute Limits? Where Do the Scaling Factors Come From? Chunky Data Exercises 170 vi

4 Chapter Twelve Avoiding Man-Made Chaos Creating Man-Made Chaos Description Is Not Analysis Charts for Each Region So What If We Combined All Six Regions? How Then Can We Compare Regions? Avoiding Man-Made Chaos A Triumph of Computation Over Common Sense Exercises 194 Chapter Thirteen Charts for Count Data Counts and Measurements Areas of Opportunity Collecting Good Count Data Charts for Counts Charts for Rates Based on Counts Some Cautions About Ratios Involving Counts Summary Exercises 211 Chapter Fourteen Traditional Charts for Count Data Charts for Binomial Counts: the np-chart Charts for Binomial Proportions: the p-chart Charts for Poisson Counts: the c-chart Charts for Poisson Rates: the u-chart So How Should You Chart Count Data? XmR Charts for Chunky Ratios Process Behavior Charts for Rare Events Summary Exercises 233 Chapter Fifteen Using Count Data Effectively Three Characteristics of Count Data The Problems of Aggregation The Difference Between Aggregations and Summaries The Power of Disaggregation Disaggregation Across Workers Disaggregation Across Problems Ratios of Aggregated Measures The Cumulative Effect of Reducing Variation Summary 260 vii

5 Chapter Sixteen Average and Range Charts Creating an Average and Range Chart What Average Charts Do Best When Does It Make Sense to Use Subgroups? Daily Department Store Sales Regional Sales by Quarter More on Peak Flow Rates Administrative Average and Range Charts Exercises 278 Chapter Seventeen Smoothing Out Seasonal Data Moving Averages Year-to-Date Plots Summary Exercises 293 Chapter Eighteen Deseasonalizing Data The Problem of Seasonal Effects Deseasonalized Values Finding Seasonal Factors Deseasonalizing the Sales Values Not All Data Show Strong Seasonality The Appliance Store Data The Claims Per Day Data Working with Seasonal Data Exercises 319 Chapter Nineteen The Learning Way Mail Order Improvement Plan, Do, Study, Act Let s Reduce Losses on Beer and Sake Sales What Should You Do Now? 344 Afterword The Germ Theory of Management 347 Appendix 365 Answers to Exercises 365 Bibliography 382 Table One Bias Correction Factors 383 Table Two Charts for Individual Values 384 Table Three Factors for use with Deseasonalized Values 385 Table Four Average and Range Charts 386 Table Five Average and Range Charts 387 Glossary of Symbols 388 Index 391 viii