Ten Data Analysis Tools You Can't Afford to be Without

Neil W. Polhemus, CTO, StatPoint Technologies, Inc.

Implementing Lean Six Sigma

Using Statgraphics

Copyright 2011 by StatPoint Technologies, Inc.

Web site: www.statgraphics.com

Lean Six Sigma

2

Lean Six Sigma

3

Lean manufacturing – focuses on reducing cost through

process optimization.

Six Sigma – focuses on meeting customer requirements and

stakeholder expectations, and improving quality by

measuring and eliminating defects.

www-935.ibm.com/services/uk/bcs/pdf

/driving_operational_innovation_using_lean_six_sigma.pdf

Statgraphics Software

4

Statgraphics Centurion XVI.I –Windows standalone application

with over 170 basic and advanced statistical methods.

Statgraphics Sigma Express 1.1 – Excel add-in with 70+

procedures covering the needs of Six Sigma green belts and

most black belts. Available for Excel 2003, 2007, 2010.*

*Pre-release evaluation version available by sending contact information to

[email protected]

Statgraphics Centurion XVI.I

5

Statgraphics Sigma Express

6

Example #1 (Define) –

Cause-and-Effect Diagram

7

Cause-and-effect diagrams (also called fishbone or Ishikawa diagrams)

illustrate the causes of specific events.

Event: defects

Major causes: material, personnel, environment, machines, …

Input – Analysis Options dialog box

creates diagram

8

Data structure for saving

9

Sigma Express also provides Excel template

10

Example #2 (Measure) – Gage Studies

11

“Gage Studies” refers to the process of evaluating measurement

processes to verify that they are capable of measuring responses well

enough to permit the use of SPC and DOE techniques.

In Statgraphics, the main procedures all follow the AIAG guidelines:

• Gage Study Setup – to create a data template.

• Analysis of Variable Data

1. Average and Range Method and ANOVA Method evaluate R&R

based on full study.

2. Range Method evaluates R&R based on short study.

3. *Gage Linearity and Accuracy evaluates bias.

• Analysis of Attribute Data

1. *Risk Assessment Method – based on consistency of appraisals

2. *Analytic Method and *Signal Theory Method – other approaches

* Not in Sigma Express product.

Statistical Model

12

2

.

2

processtmeasuremenproducttotal

22

. ilityreproducibityrepeatabilprocesstmeasuremen

Gage study setup

13

Typical gage study data file

14

Data input dialog box

15

Analysis Options

16

Tolerance = USL – LSL (distance between specification limits)

If operators measure the same parts, the structure is “crossed”.

Tables and Graphs

17

R&R Plot

18

R&R Plot for Anne_1-Carlos_3

Operators

-4

-2

0

2

4

Devia

tio

n f

rom

Avera

ge

Anne Bob Carlos

R&R Table

19

Tolerance Analysis

20

%ˆ6

100.

tolerance

processtmeasuremen

Precision-to-tolerance ratio:

Example #3 (Analyze) – Capability Analysis

21

Determines whether a process is “capable” of meeting established

specification limits.

DPMO – Defects Per Million Opportunities.

“Defect” = nonconformance to a specification.

Typical Data

22

Measured resistivity of n = 100 electronic components. USL = 500.

Capability Assessment SnapStat

23

Analysis Options

24

After Box-Cox Transformation

25

Capability Analysis Statlet®

26

Example #4 (Improve) – DOE

27

Step 1: Define Responses

Step 2 – Define Experimental Factors

28

Step 3 – Select Design

29

Step 4 – Paste to Excel Worksheet

30

Analyze Experiment

31

Tables and Graphs

32

Analysis Window

33

Paste Back to Excel

34

Example #5 (Control) – Individuals with

EWMA

35

Suggested by Stu Hunter

Data Input Dialog Box

36

X Chart – Pane Options

37

X Chart with EWMA (l = 0.2)

38

285.13

X Chart for strength

0:00 3:20 6:40 10:00 13:20 16:40

time

220

240

260

280

300

X

254.64

224.15

Note: inner zone provides 3-sigma limits for the EWMA.

Exponential Smoothing Statlet

39

Example #6 - Monte Carlo Simulation

40

Step 1

Normal (12.1, 1.9)

Step 2

Normal (7.3, 0.5)

Step 3

Exponential (0.2)

Step 4

Uniform (2.0, 6.0)

Cycle time = Step 1 + Step 2 + MAX(Step 3, Step 4)

Cycle

time

Data Input Dialog Box

41

Analysis Options

42

Results

43

Histogram of cycle time

44

Histogram

0 20 40 60 80

Cycle time

0

500

1000

1500

2000

2500

fre

qu

en

cy

More Information

Go to www.statgraphics.com

Or send e-mail to [email protected]

45