if c(version)>16 {#
8.1 codetta#
How would you display double-quotes
" "within a string?
. di `" "'
.
. di `" "quote" "'
"quote"
. di `" " " "'
" "
. di `""Strings!""'
"Strings!"
. di `"Ever heard of "for the love of the game" clause in Michael Jordans contract? "'
Ever heard of "for the love of the game" clause in Michael Jordans contract?
You are now primed for this mind twister:
di `"`label_string' `1' "`3'""'
di `" `label_string' `1' "`3'" "'
di "`label_string'"
di "`1'"
di `" "`3'" "' //token
The program below will not run after May
So you should give it a trial run in June!
By now you should be able to edit it to allow it to work beyond May
. di c(current_date) //spaces count in strings
19 May 2023
. di strpos("`c(current_date)'","May")
4
. di strpos("`c(current_date)'","May")>0
1
In the code-block below the first
if _N {is equal toif 0 {if strpos("c(current_date)’”,”May”)>0 is equal toif 1 {Just whetting your mind…
qui {
clear
cls
if _N { //background
helpful to label all variables but especially
multicategory variables as seen in q5hw1.
not only do you label the variable, you also
label values of the variable levels (alliteration).
they should be labelled if at all you wish to
efficiently produce a tidy publication-ready table1.
here we have a very condensed script to do all the above!
}
if strpos("`c(current_date)'","May")>0 { //methods
macro drop all
global github https://raw.githubusercontent.com/
global downloads jhustata/book/main/
global workdir `c(pwd)'
capture log close
log using hw1q5.log, replace
}
if _N==0 { //results
import delimited "${github}${downloads}hw1.txt"
noi di "obs: `c(N)' & vars: `c(k)'"
}
if _N { //conclusion
levelsof dx, local(dx_helper)
foreach i in `dx_helper' {
tokenize `i', p("=")
local label_string: di `"`label_string' `1' "`3'""'
}
}
if _N>0 { //acknowledgements
macro list //do this noisily!
capture label drop dx
label define dx `label_string'
split dx, p("=") generate(dx)
destring dx1, replace
label values dx1 dx //alliteration
noi tab dx1
}
if 0 { //references
1. vincent jin, your lead ta, 05/2023
2. hw1q5 ph340.600
}
log close
}
8.2 stunning#
use transplants, clear
histogram varname, normal
graph export histnormal.png, replace
. sum age,d
Age at TX
Percentiles Smallest
1% 6 0
5% 20 0
10% 29 0 Obs 6,000
25% 41 0 Sum of wgt. 6,000
50% 53 Mean 50.56117
Largest Std. dev. 15.58661
75% 62 83
90% 68 83 Variance 242.9424
95% 71 83 Skewness -.7483359
99% 77 87 Kurtosis 3.268634
8.3 transparency#
qui {
clear
if _N {
fcolor option is really neat
when overlaying histograms
one doesn't coverup the other
}
if c(N)==0 {
global url1 https://raw.githubusercontent.com/bcaffo/
global url2 ds4bme_intro/master/data/
if c(version)>17 {
di "you're good to go"
}
else {
set scheme stcolor
}
}
if _N<1 {
import delimited "${url1}${url2}oasis.csv", clear
}
if _N {
#delimit ;
hist flair_10,
legend(on
ti("Gold Lesions")
ring(0)
pos(3)
col(2)
lab(1 "Yes")
lab(2 "Simulation")
lab(3 "No")
)
xti("Flair")
yti("N",
orientation(horizontal)
)
freq
normal
fcolor(orange%40)
addplot(hist flair_20,
freq
lpattern(solid)
fcolor(midblue%50)
)
;
graph export transparency.png, replace ;
#delimit cr
}
}
8.4 scheme#
qui {
clear
if _N {
fcolor option is really neat
when overlaying histograms
one doesn't coverup the other
}
if c(N)==0 {
global url1 https://raw.githubusercontent.com/bcaffo/
global url2 ds4bme_intro/master/data/
if c(version)>17 {
set scheme s2color
}
else {
noi di "you're good to go!"
}
}
if _N<1 {
import delimited "${url1}${url2}oasis.csv", clear
}
if _N {
#delimit ;
hist flair_10,
legend(on
ti("Gold Lesions")
ring(0)
pos(3)
col(2)
lab(1 "Yes")
lab(2 "Simulation")
lab(3 "No")
)
xti("Flair")
yti("N",
orientation(horizontal)
)
freq
normal
fcolor(orange%40)
addplot(hist flair_20,
freq
lpattern(solid)
fcolor(midblue%50)
)
;
graph export transparency.png, replace ;
#delimit cr
}
}
8.5 punchline#
if c(version)>17 { //tip 4 hw3
set scheme s2color
}
else { //stcolor is default in v18
di "you're good to go!"
}
8.6 chatGPT#
Yours truly: share 20 unique conditional if code-blocks in Stata. Use a publicly available dataset to demonstrate these
Refs:
Publically available Stata datasets
Example selected by chatGPT: auto
global url https://www.stata-press.com/data/r8/
ChatGPT May12 Version:
Sure! Here are 20 unique conditional code-blocks in Stata using the publicly available “auto” dataset:
Code-block with if statement and else statement:
use ${url}auto.dta //edited to include url
local nobs=c(N)
local nvars=c(k)
webuse auto, clear
if `nobs'==c(N) & `nvars'==c(k) {
di "'webuse' is an alternative to ${url}use"
}
if foreign == 1 {
list make
}
else {
list mpg
}
This will list the make variable if the foreign variable is equal to 1, and the mpg variable otherwise.
Code-block with if statement and else-if statement:
if price < 15000 {
display "Budget car"
}
else if price < 25000 {
display "Mid-range car"
}
else {
display "Luxury car"
}
This will display a message based on the price variable, indicating whether the car is a budget car, a mid-range car, or a luxury car.
Code-block with nested if statements:
if foreign == 1 {
if mpg < 25 {
display "High-performance foreign car"
}
else {
display "Foreign car"
}
}
else {
if mpg < 20 {
display "Gas-guzzling domestic car"
}
else {
display "Domestic car"
}
}
This will display a message based on the foreign and mpg variables, indicating whether the car is a high-performance foreign car, a foreign car, a gas-guzzling domestic car, or a domestic car.
\(\vdots\)
Code-block with if statement and local macro:
local myvar "foreign"
if "`myvar'" == "foreign" {
list make
}
else {
list mpg
}
This will define a local macro called “myvar”, and then use it in an if statement to determine whether to list the make or mpg variable.
\(\vdots\)
8.7 reshape#
8.7.1 long#
Longitudinal data (time-varying covariates)
individual
time
confounders
mediators
surrogate outcomes
Hierarchical data
county
schools
public
private
charter
magnet
8.7.2 wide#
long -> wide: some loss of information + drastic increase in number of variables
Individual
confounder1 confounder2 \(\cdots\) confounderN
mediator1 mediator2 \(\cdots\) mediatorN
outcome1 outcome2 \(\cdots\) outcomeN
County
publicSch1 publicSch2 \(\cdots\) publicSchN
privateSch1 privateSch2 \(\cdots\) privateSchN
etc.
8.7.3 isid#
Quick way to determine whether a dataset is wide or long
use hw2_pra_hist.dta, clear
ds
codebook px_id
isid px_id
qui {
clear
cls
if _N {
reshape
long -> wide
then reverse
}
if c(N)==0 { //change content of macros to match location ...
global path1 ~/dropbox/2e.πρᾶξις,σ/3.acetyl.neurotrans/
global path2 big.data/big.data.usrds
global file 00_exposureR.dta
}
if _N < 1 {
use "${path1}${path2}/$file"
}
if c(N) !=0 { //using simulation to eliminate disclosure risk
noi di "obs: `c(N)', vars: `c(k)'"
g nodisclosurerisk=usrds_id+round(runiform(1,c(N)))
g ndr=round(runiform(300,365*2))
sort usrds srvc_dt
by usrds_id: egen dt=min(ndr)
replace srvc_dt=srvc_dt+dt
by usrds_id: egen id=min(nodisclosurerisk)
keep id srvc_dt insulin diur antipsy
keep id srvc_dt insulin diur antipsy
noi l id srvc_dt insulin diur antipsy in 1/30
keep in 19/26
save reshape.dta, replace
noi di "obs: `c(N)', vars: `c(k)'"
}
if _N {
//isid id
g visit=_n
keep antips* insulin diur* id visit
reshape wide antips* diur* insulin, i(id) j(visit)
}
if r(N) == 1 {
noi list
}
else {
noi di "There's more than one observation"
}
}
8.8 matcell#
use transplants, clear
tab dx gender, matcell(table)
di table[4,1] //Hypertensive Male
8.9 test.do#
takeaways from regrading hw1
presubmithomeworks in futureour team to check for only two things:
overarching
qui {controlled logfile output with no error
shouldn’t be time-consuming for student or grader
maybe do all the above via github commits? collaboratively?
i think thats where we are headed based on the perks…
not allowed to submit any hw until
presubmitclearanceencourage students to test their .do file at each step as it grows (share
repolink)as contrasted with checking a final product, which by then may have monstrously grown beyond anyones debugging skills
much more emphasis on the process, collaborative, open, decentralized repo, than a
final judgment
8.10 feedback#
feedback welcome
any system without a feedback loop is like a cancer
so please send us an email
if & when ...
8.11 suigeneris#
#delimit ; //hw1q6
table () (command result),
command(_r_b _r_ci:logistic received_kt init_age female)
nformat(%5.2f _r_b _r_ci)
sformat("(%s)" _r_ci )
cidelimiter(-) ;
collect label levels result _r_b "OR", modify ;
collect label levels result _r_ci "(95% CI)", modify ;
collect label levels
command 1 "Logistic Regression Model for Age and Sex", modify;
noi di "Question 6:" ;
noi collect preview ;
#delimit cr
8.12 myfirst#
8.12.1 unconditional#
The time it takes to write the program below
Is it dependent on, of independent of,
c(N)?How about manually estimating median (IQR) for each varname?
May this motivate you to incorporate programming into your workflow!
capture program drop myfirst
quietly program define myfirst
cls
qui ds
foreach v of varlist `r(varlist)' {
qui sum `v', d
qui local m_iqr: di %3.1f r(p50) "(" %3.1f r(p25) "-" %3.1f r(p75) ")"
if c(N) < 10 {
noi di `"The median (IQR) of "`v'" is `m_iqr' "'
}
else {
noi di "Too many values, select no more than 10 varnames"
}
}
end
use transplants, clear
myfirst
The median (IQR) of "fake_id" is 3000.5(1500.5-4500.5)
The median (IQR) of "ctr_id" is 41.0(30.0-49.0)
The median (IQR) of "transplant_date" is 19757.0(18492.0-20902.0)
The median (IQR) of "don_hgt_cm" is 172.0(163.0-180.0)
The median (IQR) of "don_wgt_kg" is 79.0(66.0-93.0)
The median (IQR) of "don_cod" is 2.0(1.0-3.0)
The median (IQR) of "don_ecd" is 0.0(0.0-0.0)
The median (IQR) of "dx" is 4.0(2.0-9.0)
The median (IQR) of "race" is 2.0(1.0-4.0)
The median (IQR) of "rec_hgt_cm" is 170.0(162.0-177.8)
The median (IQR) of "rec_wgt_kg" is 79.0(66.2-93.4)
The median (IQR) of "bmi" is 28.0(24.0-32.0)
The median (IQR) of "prev_ki" is 0.0(0.0-0.0)
The median (IQR) of "age" is 53.0(41.0-62.0)
The median (IQR) of "peak_pra" is 0.0(0.0-14.0)
The median (IQR) of "end_date" is 21868.0(21802.0-21901.0)
The median (IQR) of "died" is 0.0(0.0-0.0)
The median (IQR) of "tx_failed" is 1.0(0.0-1.0)
The median (IQR) of "wait_yrs" is 1.8(0.6-3.5)
The median (IQR) of "abo" is 3.0(1.0-4.0)
The median (IQR) of "gender" is 0.0(0.0-1.0)
The median (IQR) of "rec_hcv_antibody" is 0.0(0.0-0.0)
The median (IQR) of "rec_work" is 0.0(0.0-1.0)
The median (IQR) of "pretx_cmv" is 1.0(0.0-1.0)
The median (IQR) of "rec_education" is 2.0(2.0-3.0)
The median (IQR) of "extended_dgn" is .( .- .)
8.12.2 subcondition#
capture program drop myfirst
quietly program define myfirst
cls
qui ds
foreach v of varlist `r(varlist)' {
qui sum `v', d
qui local m_iqr: di %3.1f r(p50) "(" %3.1f r(p25) "-" %3.1f r(p75) ")"
if c(k) < 30 {
noi di `"The median (IQR) of "`v'" is `m_iqr' "'
}
else {
noi di "Too many values, select no more than 30 varnames"
}
}
end
use transplants, clear
myfirst
di c(k)
The median (IQR) of "fake_id" is 3000.5(1500.5-4500.5)
The median (IQR) of "ctr_id" is 41.0(30.0-49.0)
The median (IQR) of "transplant_date" is 19757.0(18492.0-20902.0)
The median (IQR) of "don_hgt_cm" is 172.0(163.0-180.0)
The median (IQR) of "don_wgt_kg" is 79.0(66.0-93.0)
The median (IQR) of "don_cod" is 2.0(1.0-3.0)
The median (IQR) of "don_ecd" is 0.0(0.0-0.0)
The median (IQR) of "dx" is 4.0(2.0-9.0)
The median (IQR) of "race" is 2.0(1.0-4.0)
The median (IQR) of "rec_hgt_cm" is 170.0(162.0-177.8)
The median (IQR) of "rec_wgt_kg" is 79.0(66.2-93.4)
The median (IQR) of "bmi" is 28.0(24.0-32.0)
The median (IQR) of "prev_ki" is 0.0(0.0-0.0)
The median (IQR) of "age" is 53.0(41.0-62.0)
The median (IQR) of "peak_pra" is 0.0(0.0-14.0)
The median (IQR) of "end_date" is 21868.0(21802.0-21901.0)
The median (IQR) of "died" is 0.0(0.0-0.0)
The median (IQR) of "tx_failed" is 1.0(0.0-1.0)
The median (IQR) of "wait_yrs" is 1.8(0.6-3.5)
The median (IQR) of "abo" is 3.0(1.0-4.0)
The median (IQR) of "gender" is 0.0(0.0-1.0)
The median (IQR) of "rec_hcv_antibody" is 0.0(0.0-0.0)
The median (IQR) of "rec_work" is 0.0(0.0-1.0)
The median (IQR) of "pretx_cmv" is 1.0(0.0-1.0)
The median (IQR) of "rec_education" is 2.0(2.0-3.0)
The median (IQR) of "extended_dgn" is .( .- .)
. di c(k)
26
.
end of do-file
.
8.12.3 variant#
capture program drop myfirst
quietly program define myfirst
cls
qui ds
foreach v of varlist `r(varlist)' {
qui sum `v', d
qui local m_iqr: di %3.1f r(p50) "(" %3.1f r(p25) "-" %3.1f r(p75) ")"
if c(k) > 30 {
noi di `"The median (IQR) of "`v'" is `m_iqr' "'
}
else {
noi di "Too few varnames, you can do this manually!"
exit 459
}
}
end
use transplants, clear
myfirst
di c(k)
Too few varnames, you can do this manually!
r(459);
end of do-file
r(459);
.
8.12.4 knockout#
capture program drop myfirst
quietly program define myfirst
cls
qui ds
foreach v of varlist `r(varlist)' {
qui sum `v', d
qui local m_iqr: di %3.1f r(p50) "(" %3.1f r(p25) "-" %3.1f r(p75) ")"
if c(k) > 30 {
noi di `"The median (IQR) of "`v'" is `m_iqr' "'
}
else {
noi di "Too few varnames, you can do this manually!"
//exit 459
}
}
end
use transplants, clear
myfirst
di c(k)
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
Too few varnames, you can do this manually!
. di c(k)
26
.
end of do-file
.
8.13 theline#
toe the line
cross the line
line continuation
// all these examples are programmatically equivalent
// but, aesthetically, very distinct!
// copy & paste into .do file; but don't `do`!!
super-long line of code that totally sprawls out horizontally and quite literally crosses the line if line continuation techniques aren't utilized
super-long line of code that ///
totally sprawls out horizontally ///
and quite literally crosses the line ///
if line continuation techniques aren't utilized
super-long line of code that /*
*/ totally sprawls out horizontally /*
*/ and quite literally crosses the line /*
*/ if line continuation techniques aren't utilized
#delimit ;
super-long line of code that
totally sprawls out horizontally
and quite literally crosses the line
if line continuation techniques aren't utilized
;
#delimit cr
9.14 ssc#
if 3 {//Question 3
estpost tab prev female
mat list e(colpct)
matrix define p=e(colpct)
local m_p: di %3.1f p[1,2]
local f_p: di %3.1f p[1,5]
noi di "Question 3: `m_p'% among males and `f_p'% among females have history of previous transplant."
noi di ""
}
Question 2: The median [IQR] age is 54 [42-62] among males and 51 [39-61] among females.
command estpost is unrecognized
r(199);
end of do-file
r(199);
presubmit hw
we’ll detect 3rd party programs
stata very poor with such documentation
both r & python unambiguous about libraries in each script
dead shepherd, now i find thy saw of might: whoever loved that loved not python or r?
8.15 table1_v1#
do https://raw.githubusercontent.com/jhustata/book/main/table1_options.ado
do https://raw.githubusercontent.com/jhustata/book/main/ind_translator.ado
qui {
if 0 {
}
if 1 {
cls
use 18_nhtable102feb2023.dta, clear
//g string="string"
}
if 2 {
//noi ds, has(type string)
//global string: di "`r(varlist)'"
foreach v of varlist * {
replace `v'=round(`v')
levelsof `v'
if r(r) == 2 {
noi di "binary vars: `v'"
}
else if inrange(r(r),3,30) {
noi di "categorical vars: `v'"
}
else {
noi di "continuous vars: `v'"
}
}
}
if 3 {
#delimit ;
noi table1_options,
excel("Table1_NHANES"
)
title(
"Table 1. Demographic & health characteristics of NHANES, 1988-2018"
)
by(race2
)
cont(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
)
binary(
female
smk
dm
htn
)
multi(
educ
)
foot(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
female
smk
dm
htn
educ
)
;
#delimit cr
}
}
continuous vars: seqn
binary vars: mortstat
continuous vars: permth_int
continuous vars: permth_exm
categorical vars: year
categorical vars: race
categorical vars: race2
binary vars: female
continuous vars: age
categorical vars: educ
continuous vars: income
binary vars: dm
binary vars: htn
binary vars: smk
continuous vars: sbp
continuous vars: dbp
continuous vars: bmi
categorical vars: hba1c
continuous vars: acr
categorical vars: logacr
continuous vars: glucose
categorical vars: creat
continuous vars: egfr
Table 1. Demographic & health characteristics of NHANES, 1988-2018
White Black Hispanic Asian Other
N=8234 N=5211 N=5524 N=2913 N=852
uACR, mg/g, median [IQR] 8 [5,14] 7 [4,15] 8 [5,14] 7 [5,13] 7 [4,13]
Serum Creatinine, mg/dL, median [IAR] 1 [1,1] 1 [1,1] 1 [1,1] 1 [1,1] 1 [1,1]
Income, $, median [IQR] 16579 [15,36790] 16834 [15,37404] 19012 [77,38200] 77 [15,38560] 18187 [15,37022]
BMI, kg/m2, median [IQR] 28 [24,33] 29 [25,35] 29 [26,33] 24 [22,28] 29 [24,34]
Log uACR, log(mg/g), median [IQR] 2 [2,3] 2 [1,3] 2 [2,3] 2 [2,3] 2 [1,3]
eGFR, ml/min/1.73m2, median [IAR] 98 [82,115] 113 [89,133] 110 [94,126] 109 [95,124] 105 [89,123]
Age, y, median [IQR] 51 [34,69] 49 [32,62] 46 [31,61] 44 [31,59] 40 [27,57]
SBP, mmHg, median [IQR] 120 [112,134] 126 [114,138] 120 [110,134] 118 [108,130] 120 [110,132]
HBA1c, %, median [IQR] 6 [5,6] 6 [5,6] 6 [5,6] 6 [5,6] 6 [5,6]
Glucose, mg/L, median [IQR] 9 [9,10] 9 [9,10] 10 [9,11] 9 [9,11] 9 [9,10]
Female, % 50 52 53 52 47
Smoke, % 51 41 35 23 51
Diabetes, % 14 19 17 14 15
Hypertension, % 38 44 30 24 34
Education, %
K-8 3 4 26 8 4
High school, Diploma/equivalent 34 42 39 21 31
Some college/associate 35 36 24 19 43
College graduate/above 28 19 11 52 23
acr 2.4% missing 3.2% missing 1.6% missing 1.8% missing 2.3% missing
creat 4.6% missing 10.1% missing 5.0% missing 8.1% missing 6.1% missing
income 1.6% missing 3.0% missing 3.3% missing 3.6% missing 2.0% missing
bmi 1.5% missing 1.6% missing 1.5% missing 1.0% missing 1.6% missing
logacr 2.4% missing 3.2% missing 1.6% missing 1.8% missing 2.3% missing
egfr 4.6% missing 10.1% missing 5.0% missing 8.1% missing 6.1% missing
age 0.0% missing 0.0% missing 0.0% missing 0.0% missing 0.0% missing
sbp 7.9% missing 9.0% missing 8.4% missing 9.6% missing 8.3% missing
hba1c 3.3% missing 7.5% missing 3.8% missing 6.2% missing 5.3% missing
glucose 4.6% missing 10.1% missing 5.0% missing 8.1% missing 6.1% missing
female 0.0% missing 0.0% missing 0.0% missing 0.0% missing 0.0% missing
smk 0.8% missing 2.0% missing 1.6% missing 1.1% missing 1.9% missing
dm 0.0% missing 0.0% missing 0.1% missing 0.0% missing 0.1% missing
htn 0.2% missing 0.1% missing 0.2% missing 0.1% missing 0.1% missing
educ 3.6% missing 5.6% missing 6.8% missing 4.7% missing 8.8% missing
.
end of do-file
.
8.16 table1_v2#
do https://raw.githubusercontent.com/jhustata/book/main/table1_options.ado
do https://raw.githubusercontent.com/jhustata/book/main/ind_translator.ado
qui {
if 0 {
}
if 1 {
cls
use 18_nhtable102feb2023.dta, clear
//g string="string"
}
if 2 {
//noi ds, has(type string)
//global string: di "`r(varlist)'"
foreach v of varlist * {
replace `v'=round(`v')
levelsof `v'
if r(r) == 2 {
noi di "binary vars: `v'"
}
else if inrange(r(r),3,30) {
noi di "categorical vars: `v'"
}
else {
noi di "continuous vars: `v'"
}
}
}
if 0 { //"knockout" codeblock
#delimit ;
noi table1_options,
excel("Table1_NHANES"
)
title(
"Table 1. Demographic & health characteristics of NHANES, 1988-2018"
)
by(race2
)
cont(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
)
binary(
female
smk
dm
htn
)
multi(
educ
)
foot(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
female
smk
dm
htn
educ
)
;
#delimit cr
}
}
continuous vars: seqn
binary vars: mortstat
continuous vars: permth_int
continuous vars: permth_exm
categorical vars: year
categorical vars: race
categorical vars: race2
binary vars: female
continuous vars: age
categorical vars: educ
continuous vars: income
binary vars: dm
binary vars: htn
binary vars: smk
continuous vars: sbp
continuous vars: dbp
continuous vars: bmi
categorical vars: hba1c
continuous vars: acr
categorical vars: logacr
continuous vars: glucose
categorical vars: creat
continuous vars: egfr
.
end of do-file
.
8.17 table1_v3#
do https://raw.githubusercontent.com/jhustata/book/main/table1_options.ado
do https://raw.githubusercontent.com/jhustata/book/main/ind_translator.ado
qui {
if 0 {
}
if 1 {
cls
use 18_nhtable102feb2023.dta, clear
//g string="string"
}
if 2 {
//noi ds, has(type string)
//global string: di "`r(varlist)'"
foreach v of varlist * {
replace `v'=round(`v')
levelsof `v'
if r(r) == 2 {
rename `v' b_`v'
}
else if inrange(r(r),3,30) {
rename `v' m_`v'
}
else {
rename `v' c_`v'
}
}
order c_* b_* m_*
noi list c_* in 1/10
noi list b_* in 1/10
noi list m_* in 1/10
}
if 0 { //"knockout" codeblock
#delimit ;
noi table1_options,
excel("Table1_NHANES"
)
title(
"Table 1. Demographic & health characteristics of NHANES, 1988-2018"
)
by(race2
)
cont(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
)
binary(
female
smk
dm
htn
)
multi(
educ
)
foot(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
female
smk
dm
htn
educ
)
;
#delimit cr
}
}
c_seqn c_perm~t c_perm~m c_age c_income c_sbp c_dbp c_bmi c_acr c_gluc~e c_egfr
1. -53616 109 108 85 12625 157 94 29 44 11 52
2. -53594 315 314 26 29691 124 83 37 3 9 134
3. -53593 327 326 23 25221 123 83 21 2 10 110
4. -53592 310 310 19 92696 96 67 19 31 8 141
5. -53589 6 5 81 21443 129 72 35 26 14 86
6. -53586 318 317 45 49876 130 83 53 2 10 111
7. -53585 178 177 58 . 135 65 18 2 26 110
8. -53583 337 336 51 22527 113 70 22 20 9 120
9. -53582 156 155 63 15174 116 70 21 5 9 83
10. -53581 116 114 35 56148 129 90 25 9 10 117
b_mortstat b_female b_dm b_htn b_smk
-
1. Assumed deceased 0 0 1 0
2. Assumed alive 1 0 1 0
3. Assumed alive 0 0 0 0
4. Assumed alive 1 0 0 0
5. Assumed deceased 1 0 1 1
-
6. Assumed alive 1 0 0 0
7. Assumed deceased 0 1 1 1
8. Assumed alive 1 0 0 1
9. Assumed deceased 0 0 0 1
10. Assumed deceased 0 0 0 1
m_year m_race m_race2 m_educ m_hba1c m_logacr m_creat
-
1. 1988-1998 White . High school, Diploma/equivalent 6 4 1
2. 1988-1998 White . High school, Diploma/equivalent 5 1 1
3. 1988-1998 White . College graduate/above 5 1 1
4. 1988-1998 Hispanic . High school, Diploma/equivalent 4 3 1
5. 1988-1998 White . High school, Diploma/equivalent 5 3 1
-
6. 1988-1998 White . Some college/associate 6 1 1
7. 1988-1998 Black . High school, Diploma/equivalent 8 0 1
8. 1988-1998 White . High school, Diploma/equivalent 5 3 1
9. 1988-1998 Hispanic . K-8 6 2 1
10. 1988-1998 Black . High school, Diploma/equivalent 5 2 1
.
end of do-file
.
8.18 table1_v4#
do https://raw.githubusercontent.com/jhustata/book/main/table1_options.ado
do https://raw.githubusercontent.com/jhustata/book/main/ind_translator.ado
qui {
if 0 {
}
if 1 {
cls
use 18_nhtable102feb2023.dta, clear
//g string="string"
}
if 2 {
//noi ds, has(type string)
//global string: di "`r(varlist)'"
foreach v of varlist * {
replace `v'=round(`v')
levelsof `v'
if r(r) == 2 {
rename `v' b_`v'
}
else if inrange(r(r),3,30) {
rename `v' m_`v'
}
else {
rename `v' c_`v'
}
}
order c_* b_* m_*
//noi list c_* in 1/10
//noi list b_* in 1/10
//noi list m_* in 1/10
foreach v of varlist * {
//di "`v'"
global v: di substr("`v'",3,.)
rename `v' $v
}
}
if 1 {
#delimit ;
noi table1_options,
excel("Table1_NHANES"
)
title(
"Table 1. Demographic & health characteristics of NHANES, 1988-2018"
)
/*by(race2
)*/
cont(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
)
binary(
female
smk
dm
htn
)
multi(
educ
)
foot(
acr
creat
income
bmi
logacr
egfr
age
sbp
hba1c
glucose
female
smk
dm
htn
educ
)
;
#delimit cr
}
}
Table 1. Demographic & health characteristics of NHANES, 1988-2018
uACR, mg/g, median [IQR] 7 [ 4,14]
Serum Creatinine, mg/dL, median [IAR] 1 [ 1, 1]
Income, $, median [IQR] 22846 [8193,44091]
BMI, kg/m2, median [IQR] 27 [24,32]
Log uACR, log(mg/g), median [IQR] 2 [ 1, 3]
eGFR, ml/min/1.73m2, median [IAR] 107 [88,125]
Age, y, median [IQR] 46 [30,63]
SBP, mmHg, median [IQR] 121 [111,134]
HBA1c, %, median [IQR] 5 [ 5, 6]
Glucose, mg/L, median [IQR] 9 [ 9,10]
Female, % 52
Smoke, % 46
Diabetes, % 12
Hypertension, % 31
Education, %
K-8 15
High school, Diploma/equivalent 42
Some college/associate 24
College graduate/above 19
acr: 2.0% missing
creat: 5.1% missing
income: 5.4% missing
bmi: 1.6% missing
logacr: 2.0% missing
egfr: 5.1% missing
age: 0.0% missing
sbp: 6.0% missing
hba1c: 4.0% missing
glucose: 5.1% missing
female: 0.0% missing
smk: 4.3% missing
dm: 0.1% missing
htn: 0.6% missing
educ: 5.7% missing
8.19 table1_v5#
autoencodes varnames
binary, categorical, continuous
then impliments table1_afecdvi.ado
8.20 parameters#
grades
\(\mu, \sigma\)
inferences
\(\mu =\) A
program define name
flexible
user-defined input
default if user distracted
with caveats & subconditions
mindful of c(os), c(version_real)
also of 3-party libraries
still yet to figure out the latter
\(\sigma =\) B
doesn’t illustrate knowledge of above
hasn’t presubmitted hw to test for 3-party libraries
output isn’t controlled using
qui {&noi di "..."fundamentally wrong Stata syntax
any error when script runs on our machines
8.21 ado#
update all
. update all
(contacting https://www.stata.com)
Update status
Last check for updates: 23 May 2023
New update available: none (as of 23 May 2023)
Current update level: 15 May 2023 (what's new)
Possible actions
Do nothing; all files are up to date.
.
sysdir
. sysdir
STATA: /Applications/Stata/
BASE: /Applications/Stata/ado/base/
SITE: /Applications/Stata/ado/site/
PLUS: /Users/d/Library/Application Support/Stata/ado/plus/
PERSONAL: /Users/d/Documents/Stata/ado/personal/
OLDPLACE: ~/ado/
.
ls /Applications/Stata/
. ls /Applications/Stata/
total 56
drwxrwxr-x@ 3 d admin 96 May 23 10:26 StataSE.app/
drwxrwxr-x 4 root admin 128 Apr 30 12:22 ado/
-rw-rw-r-- 1 root admin 12766 Apr 24 11:28 auto.dta
drwxrwxr-x 39 root admin 1248 Apr 24 11:28 docs/
-rw-rw-r-- 1 root admin 24 Apr 24 11:28 isstata.180
-rw-rw-r-- 1 d admin 92 Apr 26 16:13 stata.lic
-rw-r--r-- 1 d admin 3485 May 23 10:30 update.log
drwxrwxr-x 7 root admin 224 Apr 24 11:28 utilities/
.
ls /Applications/Stata/ado/base/
. ls /Applications/Stata/ado/base/
total 8
drwxrwxr-x 2 d admin 64 May 23 10:26 0/
drwxrwxr-x 2 d admin 64 May 23 10:26 1/
drwxrwxr-x 3 root admin 96 May 23 10:26 2/
drwxrwxr-x 2 d admin 64 May 23 10:26 3/
drwxrwxr-x 2 d admin 64 May 23 10:26 4/
drwxrwxr-x 2 d admin 64 May 23 10:26 5/
drwxrwxr-x 4 root admin 128 May 23 10:26 6/
drwxrwxr-x 2 d admin 64 May 23 10:26 7/
drwxrwxr-x 2 d admin 64 May 23 10:26 8/
drwxrwxr-x 2 d admin 64 May 23 10:26 9/
drwxrwxr-x 1644 root admin 52608 May 23 10:26 _/
drwxrwxr-x 213 root admin 6816 May 23 10:26 a/
drwxrwxr-x 512 root admin 16384 May 23 10:26 b/
drwxrwxr-x 795 root admin 25440 May 23 10:26 c/
drwxrwxr-x 331 root admin 10592 May 23 10:26 d/
drwxrwxr-x 381 root admin 12192 May 23 10:26 e/
drwxrwxr-x 1155 root admin 36960 May 23 10:26 f/
drwxrwxr-x 684 root admin 21888 May 23 10:26 g/
drwxrwxr-x 167 root admin 5344 May 23 10:26 h/
drwxrwxr-x 350 root admin 11200 May 23 10:26 i/
drwxrwxr-x 104 root admin 3328 May 23 10:26 j/
drwxrwxr-x 5 root admin 160 May 23 10:26 jar/
drwxrwxr-x 32 root admin 1024 May 23 10:26 k/
drwxrwxr-x 294 root admin 9408 May 23 10:26 l/
drwxrwxr-x 1349 root admin 43168 May 23 10:26 m/
drwxrwxr-x 169 root admin 5408 May 23 10:26 n/
drwxrwxr-x 85 root admin 2720 May 23 10:26 o/
drwxrwxr-x 554 root admin 17728 May 23 10:26 p/
drwxrwxr-x 4 root admin 128 May 23 10:26 py/
drwxrwxr-x 72 root admin 2304 May 23 10:26 q/
drwxrwxr-x 200 root admin 6400 May 23 10:26 r/
drwxrwxr-x 38 root admin 1216 Apr 24 11:28 resource/
drwxrwxr-x 1068 root admin 34176 May 23 10:26 s/
drwxrwxr-x 905 root admin 28960 May 23 10:26 style/
drwxrwxr-x 410 root admin 13120 May 23 10:26 t/
drwxrwxr-x 148 root admin 4736 May 23 10:26 u/
-rw-rw-r-- 1 d admin 12 May 23 10:26 update
drwxrwxr-x 154 root admin 4928 May 23 10:26 v/
drwxrwxr-x 113 root admin 3616 May 23 10:26 w/
drwxrwxr-x 371 root admin 11872 May 23 10:26 x/
drwxrwxr-x 44 root admin 1408 May 23 10:26 y/
drwxrwxr-x 81 root admin 2592 May 23 10:26 z/
ls /Users/d/Documents/Stata/ado/personal/
ls: /Users/d/Documents/Stata/ado/personal/: No such file or directory
.
Probably a folder Stata expects you to create and utilize for 3rd-party .ado files
Every student should be aware of their location
Install libraries via
ssccommand in .do file script to avoid crashing on collaborators machineThis, of course, also applies to hw since it simulates a collaboration
which coefplot
. which coefplot
/Users/d/Library/Application Support/Stata/ado/plus/c/coefplot.ado
*! version 1.8.6 22feb2023 Ben Jann
.
Above
coefplotcommand is, clearly, from a third-partyMaybe write a “screener” script which tests every command?
global commandlist if g sum coefplot tokenize //built-in commands appear blue
local n: di wordcount("$commandlist")
forval i=1/`n' {
local command: di word("$commandlist",`i')
which "`command'"
}
. global commandlist if g sum coefplot tokenize //built-in commands appear blue
. local n: di wordcount("$commandlist")
. forval i=1/`n' {
2. local command: di word("$commandlist",`i')
3. which "`command'"
4. }
built-in command: if
built-in command: generate
built-in command: summarize
/Users/d/Library/Application Support/Stata/ado/plus/c/coefplot.ado
*! version 1.8.6 22feb2023 Ben Jann
built-in command: tokenize
.
end of do-file
.
In the above example we can already see that coefplot appears white
global commandlist if g sum coefplot tokenize estpost //built-in commands appear blue
local n: di wordcount("$commandlist")
forval i=1/`n' {
local command: di word("$commandlist",`i')
which "`command'"
}
. global commandlist if g sum coefplot tokenize estpost //built-in commands appear blue
. local n: di wordcount("$commandlist")
. forval i=1/`n' {
2. local command: di word("$commandlist",`i')
3. which "`command'"
4. }
built-in command: if
built-in command: generate
built-in command: summarize
/Users/d/Library/Application Support/Stata/ado/plus/c/coefplot.ado
*! version 1.8.6 22feb2023 Ben Jann
built-in command: tokenize
command estpost not found as either built-in or ado-file
r(111);
end of do-file
r(111);
ssc install estpost
ssc install coefplot
. ssc install estpost
ssc install: "estpost" not found at SSC, type search estpost
(To find all packages at SSC that start with e, type ssc describe e)
r(601);
. ssc install coefplot
checking coefplot consistency and verifying not already installed...
all files already exist and are up to date.
.
8.22 dofile#
qui {
cls
clear
if _N { //background
1. done in the spirit of a workflow that incorporates collaboration
2. you, the student, programmer, and collaborator must be mindful
3. mindful of the library of third-party programs you've installed
4. these may not necessarily be installed on your collaborators' machine
5. as we grade your hws we'll act like frustrated collaborators if...
6. we run your .do file and get errors on our machine
7. so we are not permanently incorporating a codeblock to address this
8. edit ${workdir} before you share your code with collaborators!!!!!!!!
}
if c(N)==_N { //method
#delimit ;
global commandlist
qui
cls
clear
if
coefplot
help
tokenize
pwd
program
capture
twoway
local
;
#delimit cr
local n: di wordcount("$commandlist")
forval i=1/`n' {
local command: di word("$commandlist",`i')
noi which "`command'"
}
if c(os) == "Windows" { //paths with back "\" slashes
global workdir `c(pwd)'
}
else { //paths with forward "/" slashes
global workdir `c(pwd)'
}
capture log close
log using statalibrary.log, replace
set more off
}
if _N < 1 { //results
//etc.
}
}
built-in command: quietly
built-in command: cls
/Applications/Stata/ado/base/c/clear.ado
*! version 4.4.0 26feb2021
built-in command: if
/Users/d/Library/Application Support/Stata/ado/plus/c/coefplot.ado
*! version 1.8.6 22feb2023 Ben Jann
/Applications/Stata/ado/base/h/help.ado
*! version 1.0.1 18sep2019
built-in command: tokenize
built-in command: pwd
built-in command: program
built-in command: capture
/Applications/Stata/ado/base/t/twoway.ado
*! version 1.2.1 22jun2012
built-in command: local
.
end of do-file
.