ejecting data from the mosaic package and putting it into mosaicData

[rpruim]

For the sake of keeping CRAN happy.

[rpruim]

I'm in the process of making sure I have

if (require(mosaicData))

in tall the required places.

Also, I'm putting NHANES in a package all by itself.

[rpruim]

using mosaicData:: in the tests so that they work...

[dtkaplan]

While we're at it, let's change NHANES a bit to delete some of the inessential variables.

Here's an Rda file and the corresponding Rd. I was calling the file NHANEDCF, but it should be renamed simply NHANES.

-Danny

On Thu, Aug 21, 2014 at 11:01 PM, Randall Pruim notifications@github.com wrote:

I'm in the process of making sure I have

if (require(mosaicData))

in tall the required places.

Also, I'm putting NHANES in a package all by itself.

— Reply to this email directly or view it on GitHub https://github.com/rpruim/mosaic/issues/369#issuecomment-53020126.

[rpruim]

I don't see an Rda file. But it would suffice to provide a select() command that drops or keeps the columns desired.

For documentation, it would be better to have an R file with the appropriate roxygen markup. See datasets.R

[rpruim]

Which of these are you proposing we drop?

names(NHANES)
 [1] "seqn"        "sex"         "age"         "pregnant"    "ethnicity"   "death"      
 [7] "followup"    "smoker"      "diabetic"    "height"      "weight"      "waist"      
[13] "wci"         "bmi"         "ptfat"       "tfat"        "lfat"        "llean"      
[19] "lbmi"        "fbmi"        "bbmi"        "pfat"        "bmd"         "fmhm_other" 
[25] "hdl"         "chol"        "bps"         "bpd"         "income"      "zheight"    
[31] "zweight"     "zwaist"      "zwci"        "zbmi"        "zptfat"      "ztfat"      
[37] "zlfat"       "zllean"      "zlbmi"       "zfbmi"       "zbbmi"       "zpfat"      
[43] "zbmd"        "zfmhm_other" "zhdl"        "zchol"       "zbps"        "zbpd"       
[49] "zincome"     "pop_weight"  "psu"         "stratum"     "zwh"  

I'm making a final spin through vignettes, checking and double checking. I want to submit this VERY soon. So I won't wait for NHANES edits if we can't do them quickly.

[nicholasjhorton]

I appreciate your prioritizing this.

Nick

On Aug 22, 2014, at 9:03 AM, Randall Pruim notifications@github.com wrote:

Which of these are you proposing we drop?

names(NHANES) [1] "seqn" "sex" "age" "pregnant" "ethnicity" "death"
[7] "followup" "smoker" "diabetic" "height" "weight" "waist"
[13] "wci" "bmi" "ptfat" "tfat" "lfat" "llean"
[19] "lbmi" "fbmi" "bbmi" "pfat" "bmd" "fmhm_other" [25] "hdl" "chol" "bps" "bpd" "income" "zheight"
[31] "zweight" "zwaist" "zwci" "zbmi" "zptfat" "ztfat"
[37] "zlfat" "zllean" "zlbmi" "zfbmi" "zbbmi" "zpfat"
[43] "zbmd" "zfmhm_other" "zhdl" "zchol" "zbps" "zbpd"
[49] "zincome" "pop_weight" "psu" "stratum" "zwh"

I'm making a final spin through vignettes, checking and double checking. I want to submit this VERY soon. So I won't wait for NHANES edits if we can't do them quickly.

— Reply to this email directly or view it on GitHub.

[rpruim]

I'd like Nick to weigh in on column dropping, since I think he uses this data and we don't want to break things he is relying upon.

[nicholasjhorton]

I have no strong feelings (but I don't use the scored variables).

All the best,

Nick

On Aug 22, 2014, at 9:04 AM, Randall Pruim notifications@github.com wrote:

I'd like Nick to weigh in on column dropping, since I think he uses this data and we don't want to break things he is relying upon.

— Reply to this email directly or view it on GitHub.

[rpruim]

OK. I'll remove the scaled variables. That will make it roughly half the size and easier to work with.

Makes we wonder about adding a zscore function:

zscore <- function( x, na.rm=getOptions("na.rm", FALSE) ) {
  ( x - mean(x, na.rm=na.rm)) / sd(x, na.rm=na.rm)
}

That's easy enough, I think I'll do it.

[rpruim]

zscores have been removed from NHANES.

[nicholasjhorton]

Isn't that what "scale()" does?

Nick

On Aug 22, 2014, at 9:32 AM, Randall Pruim notifications@github.com wrote:

OK. I'll remove the scaled variables. That will make it roughly half the size and easier to work with.

Makes we wonder about adding a zscore function:

zscore <- function( x, na.rm=getOptions("na.rm", FALSE) ) {

( x - mean(x, na.rm=na.rm)) / sd(x, na.rm=na.rm) } That's easy enough, I think I'll do it.

— Reply to this email directly or view it on GitHub.