I have a box plot with multiple levels d$a + d$b
d = data.frame(value = c(1,2,3,100), a = c("A","A","B","A"), b = c("C","D","C","C") )
boxplot(d$value ~ d$a + d$b, horizontal = TRUE)
when you run that code you will see that the B.D combination still shows up but it is empty. How do I remove it?
This is just a toy example. In reality I have 40+ combinations and do not want to remove the blank one by hand.
You can first use interaction (together with its drop argument) to create a new column of your data.frame, then plot it:
d <- data.frame(value = c(1,2,3,100), a = c("A","A","B","A"), b = c("C","D","C","C"))
d <- within(d, interaction <- interaction(a, b, drop = TRUE))
boxplot(value ~ interaction, data = d, horizontal = TRUE)
Related
I have a dataset that contains both numeric and categorical values. I am trying to create box plots to visually identify outliers for each numeric column in my dataset. The below code works to do this, but it is very clunky and I would not want to use this code with even more variables. I am looking for a way to use a loop to create box plots using a loop in R.
Here is the clunky code that works without a loop:
#Using Boxplots, check for outliers in each in each float or integer value column.
b <-boxplot(df$item1, main = 'item1')
b <-boxplot(df$item2, main = 'item2')
b <-boxplot(df$item3, main = 'item3')
b <-boxplot(df$item4, main = 'item4')
b <-boxplot(df$item5, main = 'item5')
b <-boxplot(df$item6, main = 'item6')
b <-boxplot(df$item7, main = 'item7')
b <-boxplot(df$item8, main = 'item8')
b <-boxplot(df$item9, main = 'item9')
b <-boxplot(df$item10, main = 'item10')
b <-boxplot(df$item11, main = 'item11')
b <-boxplot(df$item12, main = 'item12')
b <-boxplot(df$item13, main = 'item13')
b <-boxplot(df$item14, main = 'item14')
b <-boxplot(df$item15, main = 'item15')
b <-boxplot(df$item16, main = 'item16')
In python the code would be:
outliers = ['Item1', 'Item2', 'Item3', 'Item4', 'Item5', 'Item6', 'Item7', 'Item8', 'Item9', 'Item10', 'Item11', 'Item12', 'Item13', 'Item14', 'Item15', 'Item16']
i=0
while i < len(outliers):
sns.boxplot(x = outliers[i], data = df)
plt.show()
i = i + 1
(I am looking for something similar in R!)
Thank you!
Using a for loop to loop over the columns and a minimal reprex based on mtcars you could do
outliers <- c("mpg", "hp")
for (i in outliers) {
boxplot(mtcars[i], main = i)
}
Put simply, I want to color outliers, but only if they belong to specific category, i.e. I want
boxplot(mydata[,2:3], col=c("chartreuse","gold"), outcol="red")
but red only for those elements for which mydata[,1] is M .
It appears that outcol only specifies one color per variable (box). However, you can use points to overplot individual points any way that you want. You need to figure out the relevant x and y coordinates to use for plotting. When you make a boxplot with a statement like boxplot(mydata[,2:3]) the first variable (column 2) is plotted at x=1 and the second variable (column 3) is plotted at x=2. By capturing the return value of boxplot you can figure out the y values. Since you do not provide any data, I will illustrate with randomly generated data.
## Data
set.seed(42)
NumPts = 400
a = rnorm(NumPts)
b = rnorm(NumPts)
c = rnorm(NumPts)
CAT = sample(c("M", "N"), NumPts, replace=T)
mydata = data.frame(a,b,c, CAT)
## Find outliers
BP = boxplot(mydata[,2:3], col=c("chartreuse","gold"))
OUT2 = which(mydata[,2] %in% BP$out)
OUT3 = which(mydata[,3] %in% BP$out)
## Find outliers with category == M
M_OUT2 = OUT2[which(mydata$CAT[OUT2] == "M")]
M_OUT3 = OUT3[which(mydata$CAT[OUT3] == "M")]
## Plot desired points
points(rep(1, length(M_OUT2)),mydata[M_OUT2, 2], col="red")
points(rep(2, length(M_OUT3)),mydata[M_OUT3, 3], col="red")
I have a dataset with 99 observations and I need to create boxplots for ones with a specific string in them. However, when I run this code I get 57 of the exact same plots from the original function instead of the loop. I was wondering how to prevent the plots from being overwritten but still create all 57. Here is the code and a picture of the plot.
Thanks!
Boxplot Format
#starting boxplot function
myboxplot <- function(mydata=ivf_dataset, myexposure =
"ART_CURRENT", myoutcome = "MEG3_DMR_mean")
{bp <- ggplot(ivf_dataset, aes(ART_CURRENT, MEG3_DMR_mean))
bp <- bp + geom_boxplot(aes(group =ART_CURRENT))
}
#pulling out variables needed for plots
outcomes = names(ivf_dataset)[grep("_DMR_", names(ivf_dataset),
ignore.case = T)]
#creating loop for 57 boxplots
allplots <- list()
for (i in seq_along(outcomes))
{
allplots[[i]]<- myboxplot (myexposure = "ART_CURRENT", myoutcome =
outcomes[i])
}
allplots
I recommend reading about standard and non-standard evaluation and how this works with the tidyverse. Here are some links
http://adv-r.had.co.nz/Functions.html#function-arguments
http://adv-r.had.co.nz/Computing-on-the-language.html
I also found this useful
https://rstudio-pubs-static.s3.amazonaws.com/97970_465837f898094848b293e3988a1328c6.html
Also, you need to produce an example so that it is possible to replicate your problem. Here is the data that I created.
df <- data.frame(label = rep(c("a","b","c"), 5),
x = rnorm(15),
y = rnorm(15),
x2 = rnorm(15, 10),
y2 = rnorm(15, 5))
I kept most of your code the same and only changed what needed to be changed.
myboxplot2 <- function(mydata = df, myexposure, myoutcome){
bp <- ggplot(mydata, aes_(as.name(myexposure), as.name(myoutcome))) +
geom_boxplot()
print(bp)
}
myboxplot2(myexposure = "label", myoutcome = "y")
Because aes() uses non-standard evaluation, you need to use aes_(). Again, read the links above.
Here I am getting all the columns that start with x. I am assuming that your code gets the columns that you want.
outcomes <- names(df)[grep("^x", names(df), ignore.case = TRUE)]
Here I am looping through in the same way that you did. I am only storing the plot object though.
allplots <- list()
for (i in seq_along(outcomes)){
allplots[[i]]<- myboxplot2(myexposure = "label", myoutcome = outcomes[i])$plot
}
allplots
I just discovered the power of plyr frequency table with several variables in R
and I am still struggling to understand how it works and I hope some here can help me.
I would like to create a table (data frame) in which I can combine frequencies and summary stats but without hard-coding the values.
Here an example dataset
require(datasets)
d1 <- sleep
# I classify the variable extra to calculate the frequencies
extraClassified <- cut(d1$extra, breaks = 3, labels = c('low', 'medium', 'high') )
d1 <- data.frame(d1, extraClassified)
The results I am looking for should look like that :
require(plyr)
ddply(d1, "group", summarise,
All = length(ID),
nLow = sum(extraClassified == "low"),
nMedium = sum(extraClassified == "medium"),
nHigh = sum(extraClassified == "high"),
PctLow = round(sum(extraClassified == "low")/ length(ID), digits = 1),
PctMedium = round(sum(extraClassified == "medium")/ length(ID), digits = 1),
PctHigh = round(sum(extraClassified == "high")/ length(ID), digits = 1),
xmean = round(mean(extra), digits = 1),
xsd = round(sd(extra), digits = 1))
My question: how can I do this without hard-coding the values?
For the records:
I tried this code, but it does not work
ddply (d1, "group",
function(i) c(table(i$extraClassified),
prop.table(as.character(i$extraClassified))),
)
Thanks in advance
Here's an example to get you started:
foo <- function(x,colfac,colval){
tbl <- table(x[,colfac])
res <- cbind(n = nrow(x),t(tbl),t(prop.table(tbl)))
colnames(res)[5:7] <- paste(colnames(res)[5:7],"Pct",sep = "")
res <- as.data.frame(res)
res$mn <- mean(x[,colval])
res$sd <- sd(x[,colval])
res
}
ddply(d1,.(group),foo,colfac = "extraClassified",colval = "extra")
Don't take anything in that function foo as gospel. I just wrote that off the top of my head. Surely improvements/modifications are possible, but at least it's something to start with.
Thanks to Joran.
I slighlty modified your function to make it more generic (without reference to the position of the variables) .
require(plyr)
foo <- function(x,colfac,colval)
{
# table with frequencies
tbl <- table(x[,colfac])
# table with percentages
tblpct <- t(prop.table(tbl))
colnames( tblpct) <- paste(colnames(t(tbl)), 'Pct', sep = '')
# put the first part together
res <- cbind(n = nrow(x), t(tbl), tblpct)
res <- as.data.frame(res)
# add summary statistics
res$mn <- mean(x[,colval])
res$sd <- sd(x[,colval])
res
}
ddply(d1,.(group),foo,colfac = "extraClassified",colval = "extra")
and it works !!!
P.S : I still do not understand what (group) stands for but
This is a follow up question from R: t-test over all columns
Suppose I have a huge data set, and then I created numerous subsets based on certain conditions. The subsets should have the same number of columns. Then I want to do t-test on two subsets at a time (outer loop) and then for each combination of subsets go through all columns one column at a time (inner loop).
Here is what I have come up with based on previous answer. This one stops with an error.
C <- c("c1","c1","c1","c1","c1",
"c2","c2","c2","c2","c2",
"c3","c3","c3","c3","c3",
"c4","c4","c4","c4","c4",
"c5","c5","c5","c5","c5",
"c6","c6","c6","c6","c6",
"c7","c7","c7","c7","c7",
"c8","c8","c8","c8","c8",
"c9","c9","c9","c9","c9",
"c10","c10","c10","c10","c10")
X <- rnorm(n=50, mean = 10, sd = 5)
Y <- rnorm(n=50, mean = 15, sd = 6)
Z <- rnorm(n=50, mean = 20, sd = 5)
Data <- data.frame(C, X, Y, Z)
Data.c1 = subset(Data, C == "c1",select=X:Z)
Data.c2 = subset(Data, C == "c2",select=X:Z)
Data.c3 = subset(Data, C == "c3",select=X:Z)
Data.c4 = subset(Data, C == "c4",select=X:Z)
Data.c5 = subset(Data, C == "c5",select=X:Z)
Data.Subsets = c("Data.c1",
"Data.c2",
"Data.c3",
"Data.c4",
"Data.c5")
library(plyr)
combo1 <- combn(length(Data.Subsets),1)
adply(combo1, 1, function(x) {
combo2 <- combn(ncol(Data.Subsets[x]),2)
adply(combo2, 2, function(y) {
test <- t.test( Data.Subsets[x][, y[1]], Data.Subsets[x][, y[2]], na.rm=TRUE)
out <- data.frame("Subset" = rownames(Data.Subsets[x]),
, "Row" = colnames(x)[y[1]]
, "Column" = colnames(x[y[2]])
, "t.value" = round(test$statistic,3)
, "df"= test$parameter
, "p.value" = round(test$p.value, 3)
)
return(out)
} )
} )
First of all, you can more easily define you dataset using gl, and by avoiding creating individual variables for the columns.
Data <- data.frame(
C = gl(10, 5, labels = paste("c", 1:10, sep = "")),
X = rnorm(n = 50, mean = 10, sd = 5),
Y = rnorm(n = 50, mean = 15, sd = 6),
Z = rnorm(n = 50, mean = 20, sd = 5)
)
Convert this to "long" format using melt from the reshape package. (You can also use the base reshape function.)
longData <- melt(Data, id.vars = "C")
Now Use pairwise.t.test to compute t tests on all pairs of X/Y/Z for for each level of C.
with(longData, pairwise.t.test(value, interaction(C, variable)))
Note that it is important to use pairwise.t.test rather than just lots of individual calls to t.test because you need to adjust your p values if you run lots of tests. (See, e.g., xkcd for explanation.)
In general, pairwise t tests are inferior to a regression so be careful about their usage.
You can use get(Data.subset[x]) which will pick out the relevant data frame. But I don't think this should be necessary.
Explicitly subsetting that many times shoudn't be necessry either. You could create them using something like
conditions = c("c1", "c2", "c3", "c4", "c5")
dfs <- lapply(conditions, function(x){subset(Data, C==x, select=X:Z)})
That should (didn't test it) return a list of data frames each subseted on the various conditions you passed it.
However it would be a much better idea as #Richie Cotton points out, to reshape your data frame and use pairwise t tests.
I should point out that doing this many t-tests doesn't seem wise. Even after correction for multiple testing, be it FDR, permutation or otherwise. It would be better to try and figure out if you can use an anova of some sort as they are used for almost exactly this purpose.