emilopezcano/sixsigma

SixSigma R package

sixsigma's Issues

Use ss.col when calling .ss.repCanvas in ss.pmap

Propagate colors at:

SixSigma/R/ss.pMap.R

Line 92 in 12a429d

.ss.prepCanvas(main,sub)

Merge master into develop

For some time, changes were done in master and now are not in develop.

After nice contributions of @k-tatgenhorst and in view to further contributions, I go back to the flow:

1. Open issue
2. Create issue branch on develop (pattern: number_issue/short_description
3. Commits to branch on develop
4. Review and merge branch with develop
5. Test on develop
6. Merge with master
7. Remove issue branch

So now it is time to merge master into develop

Plot Backgrounds on Capabilities Analysis

I submitted an update for ss.ca.r which gives control over the colors.
The plots on the left continue to have a white background when the background color is changed.
It is not horrible, but not optimal. I believe it is related to using viewport without defining a rectange, so
it either needs defined rectangles or perhaps a background can be set in the actual plot.

As this was only a minor aesthetic, I chose to submit as is and address later. If anyone reading this knows, please feel
free to tweak it.

Improve package structure

Some improvements are needed to follow best practices:

• Data as .rda
• Check tests to pass CHECK

Avoid spurious warning in ss.rr

Avoid this warning

Frio the book's example 5.5:

my.rr <- ss.rr(var = voltage, part = battery, appr = voltmeter, data = batteries, main = "Six Sigma Gage R&R Measure", sub = "Batteries Project MSA")

The output produces this warning:

Warning message:
In widths.x[pos.widths[[nm]]] <- widths.settings[[nm]] * widths.defaults[[nm]]$x :
  number of items to replace is not a multiple of replacement length

Vignette Shewhart Constants

Make work as it is said in the book:

vignette(package = "Six Sigma", topic = "Shewhart Constants")

(with the space)

R&R for attributes

New function for R&R analysis for attributes
Write new function
Write unit test

Asun & Javier as contributors (function and DESCRIPTION)

Source: https://bookdown.org/asun_mayoral/book-l6s/dmaic2-medir.html#análisis-del-sistema-de-medida and script linked therein.

Implement R&R Type 1 analysis

Implement a new function for R&R Type 1 analysis, also cover the possibility to use only one spec limit.

Implement R&R analysis for unbalanced designs

Implement R&R analysis for unbalanced designs

Modify the ss.rr() function to switch the execution depending on the balance of the design.

Check for the standards and implement the correct method to work with unbalanced designs. The method for balanced ones is already implemented and working.

Use the value of appr, not the string "appr"

SixSigma/R/ss.rr.R

Line 396 in f25d641

main = expression(bold(bar(x)*" Chart by "*appr)),

Check for the purpose of errorTerm param in ss.rr function

Check for the purpose of the param errorTerm in the MSA analysis, within ss.rr() function.

Implement functionality if it was the purpose of errorTerm.

Frame color is not being passed in pMap

ss.pMap constructs a list of colors currently five elements long. Those elements should be passed to the .ss.prepCanvas function but the call omits that variable. ss.col elements 1-4 are hardcoded into the prepCanvas function (as list elements) so they still work as expected, but where element 5 should be, it is hard coded as white.

The changes that I made for my previous issue actually rectify this, but for purposes of historical accuracy - I wanted to submit an issue highlighting this.

Implement R&R analysis for designs with no replications

Implement R&R analysis for designs with no replications

Modify the ss.rr() function to switch the execution and work with the correct model.

Check for the standards for the correct method. Right now the function returns an error and inform the user that designs with no replications are not implemented.

Updates to ggplot break ss.study.ca

It looks like binwidth is no longer a valid argument for stat_density

> ss.study.ca(ss.data.ca$Volume, rnorm(40, 753, 3), 
+             LSL = 740, USL = 760, T = 750, alpha = 0.05, 
+             f.sub = "Winery Project")

> traceback()
4: stop("Unknown parameters: ", paste(extra, collapse = ", "), call. = FALSE)
3: layer(data = data, mapping = mapping, stat = StatDensity, geom = geom, 
       position = position, show.legend = show.legend, inherit.aes = inherit.aes, 
       params = list(adjust = adjust, kernel = kernel, trim = trim, 
           na.rm = na.rm, ...))
2: stat_density(geom = "path", data = ggdataLT, position = "identity", 
       binwidth = binwLT) at #119
1: ss.study.ca(ss.data.ca$Volume, rnorm(40, 753, 3), LSL = 740, 
       USL = 760, T = 750, alpha = 0.05, f.sub = "Winery Project")

Improvements in ss.rr plots

User request for ss.rr:

But how can I change the titles in the subplots.
Most automatically adopt the name of the column that is specified such as appr = operator (in my case). But for the xbar > > chart it uses “appr”. Also, how could I change the position of the plots?

Further customization in ss.pMap

ss.Pmap does not include methods to change the following attributes:

Text Color
Background Color

These would be useful for organizations with branding requirements.

Additionally, document output size.

Are the elements drawn proportionally to the document size? ie.. if I add many more steps, will they be drawn smaller or overflow the document? <- I am guessing that this would probably need it's own issue if it does not currently scale.

Error when no variation by operator in rr

SixSigma/R/ss.rr.R

Line 582 in 5ec13ee

glimits <- c(min(range(data.xrange[[var]])[1], rlimits[1]),

When there is no variation by appraisers, ranges are 0 and the function fails to set the y-axis limits.

Thanks to Gilles CHASSAT

Implement S4 classes

S4 objects

spc

@DaTa (data.frame)
@controlchart: list of (type of chart). Each one a list: UCL, LCL, CL, sigma, nsigmas, points
@ctq (chart) If NULL -> rownames
@hlines
@flimits
@newrows
@controlledrows
@order
@special: list of: out, ...
@zones

@capability: list of: Target, USL, LSL, Indices (cp, cpk, cpm, ppk, ...)

When creating object: calculate limits

plot method: mark zones?, type, nsigmas,

sqc

For acceptance sampling

Methods

• Create object
• Fill @controlchart
• generic plot function

Avoiding implicit overwrite of "show.signif.stars"

This package is fantastic !
But I have one thing that I would like to ask.

It would be nice if ss.rr does not implicitly overwrite "show.signif.stars".

Here is my suggestion patch.

Thank you.

ss-rr-signifstars-patch.txt