R Functions

Custom Figures Misc Functions Two-Stage Designs Randomization

Miscellaneous Functions

groupSum computes and displays a summary of a numeric vector v, optionally by grouped by g. Missing group IDs are treated as a separate group.

data(iris)
iris$Species[seq(1,150,length=20)] <- NA ## for demonstration purposes.

groupSum(v=Sepal.Length, g=Species, data=iris, select_stats=c('-SE')


             N Min   Q1 Med   Q3 Max Mean    SD
setosa      43 4.3 4.80 5.0 5.20 5.8 4.98 0.353
versicolor  44 4.9 5.57 5.8 6.30 7.0 5.91 0.528
virginica   43 4.9 6.25 6.5 6.85 7.9 6.57 0.627
NA.         20 4.6 5.33 5.9 6.55 7.7 5.99 0.842
Combined   150 4.3 5.10 5.8 6.40 7.9 5.84 0.828

editRedCapData cleans the data.frame created by REDCap's .r file.
- Delete all the numeric vectors when corresponding factor vectors are available.
- Delete .factor from the variable names.
- Change . to _ in the variable names (optional)
```
d <- editRedCapData(d, changeNames=TRUE)
```
multTime multiplies a time value, lap, by multiplier.
This function answers questions such as: If I run 4:12/km for 42.195km, how long will it take?
- The lap argument can be specified in "HH:MM:SS" or "MM:SS" format (character string) or MM.SS format (numeric).
```
multTime(lap=4.12, multiplier=42.195)
# multTime(lap='4:12', multiplier=42.195) # Same

[1] "2:57:13"
```
addTime sums time values from the vector times, which can be specified in "HH:MM:SS" or "MM:SS" format (character string) or MM.SS format (numeric).
- The outputs of multTime and addTime can be used as inputs for this function. For example, to answer If I run 4:02/km for 5km, then 3:55/km for 7km, what is the total time?
```
addTime( c(multTime("4:02",5), multTime(3.55,7)))

[1] "47:35"
```

BinPower finds the critical value(s) for one-sample Binomial test.

Xcr in the output is the critical value (inclusive). H₀ is rejected if X is Xcr_high or higher, or X is Xcr_low or lower.

BinPower(n=50, p0=0.2, p1=0.4, alp=0.05, two.sided=FALSE)

   n  p0  p1  alp two.sided Xcr_low Xcr_high  typeI Power
1 50 0.2 0.4 0.05     FALSE      NA       16 0.0308 0.904

For two-sided test, type I error rate, alp, is divided by 2.

BinPower(n=50, p0=0.2, p1=0.4, alp=0.05, two.sided=TRUE)

   n  p0  p1  alp two.sided Xcr_low Xcr_high  typeI Power
1 50 0.2 0.4 0.05      TRUE       4       17 0.0329 0.844

Beta distribution. These four functions are used for Beta distributions. a and b are the two parameters.
- BetaMS(a, b) calculates the mean, variance, and standard deviation of a Beta(a, b) distribution.
- plotBetaDis(a, b) plots the density function given a and b.
- BetaParameters(m, v) solves for a and b given the mean and variance of the distribution. I use this function when I need to simulate probabilities.
- BetaCI(a, b, x, n, p) computes a 100*p% credible interval given Beta(a, b) as the prior distribution and "x successes out n" as the data.
```
BetaMS(a=29, b=11)

   a  b  mean     var     sd
1 29 11 0.725 0.00486 0.0697
```
```
plotBetaDis(a=4, b=21, xlim=c(0, 0.5), yaxs='i')
```
```
BetaParameters(m=0.4, v=0.02, PLOT=FALSE)

[1] 4.4 6.6
```
```
BetaCI(a=4.4, b=6.6, x=8, n=15, p=0.95)

  prior_a prior_b successes trials cred_level lower median upper
1     4.4     6.6         8     15        95% 0.292  0.476 0.665
```
Survival Parameters. These functions convert survival parameters assuming a constant hazard, as in the exponential survival model.
- ps2hr(prop, Time) converts proportion surviving (at Time = Time) to hazard rate.
- ms2hr(ms) converts median survival to hazard rate.
- hr2ms(hr) converts hazard rate to median survival.
- ps2ms(prop, Time) converts proportion surviving (at Time = Time) to median survival.
- ms2ps(ms, Time) converts median survival to proportion surviving.
- hr2ps(hr, Time) converts hazard rate to proportion surviving.
What is the one-year survival proportion if the median survival is 5 months?
```
ms2ps(ms=5, Time=12)

[1] 0.189
```
myKindaDay takes a regular R 'date' element and changes it to "yyyy/mm/dd" with no leading zeros (which is most common in Japan).
```
as.Date('1999-6-20')

[1] "1999-06-20"
```
```
myKindaDay(as.Date('1999-6-20'))

[1] "1999/6/20"
```