How to create histograms for each unique combination of levels from two factors?

Question

I cannot figure out how to use a loop to plot one histogram for each unique combination of levels from TWO factors.

Here is my data: https://www.dropbox.com/sh/exsjhu23fnpwf4r/AABvitLBN1nRMpXcyYMVIOIDa?dl=0

# perhaps need to have factors
df$freq <- as.factor(df$freq)
df$time <- as.factor(df$time)

I learned how to use a loop to plot histograms for ONE factor levels:

# space for plots  
windows(width=19, height=10) 
par(las=1, cex.lab=0.75, cex.axis=0.6, bty="n", mgp=c(1, 0.6, 0),
    oma=c(2, 4, 2, 0) + 0.1, mar=c(4, 0, 3, 3) + 0.1)
a <- layout(matrix(c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,
                     18, 19, 20, 21), nrow=3, ncol=7, byrow=T))
layout.show(a)

# loop 
for (i in 1:length(unique(df$freq))) {
  value <- subset(df, freq == unique (df$freq)[i])
  hist(value$thr, main=paste0("freq: ", unique(df$freq)[i]))
}

I tried variations of this loop for TWO factors but that unfortunately does not work:

for (i in 1:length(unique(df[c("freq", "time")]))) {
  value <- subset(df, freq == unique (df$freq)[i] & time == unique(df$time)[i])
  hist(value$thr, main=paste0("freq: ", unique(df$freq)[i]))
}

I would also like to learn how to label each histogram based on the levels of TWO factors (not just one)...

Answer 1

It's more convenient to use by here.

For the titles we prefer characters to factors.

df1[c("freq", "time")] <- lapply(df1[c("freq", "time")], as.character)

Then open windows,

windows(width=19, height=10) 
par(las=1, cex.lab=0.75, cex.axis=0.6, bty="n", mgp=c(1, 0.6, 0), 
    oma=c(2, 4, 2, 0) + 0.1, mar=c(4, 0, 3, 3) + 0.1)
a <- layout(matrix(1:21, 3, 7))
layout.show(a)

and plot.

by(df1, df1[c("freq", "time")], function(x) 
  hist(x$thr, main=paste("freq:", paste(x[1, c(1, 3)], collapse=","))))

Result

Edit

To get the specific order we probably have to do some more stuff.

df1[c("freq", "time")] <- lapply(df1[c("freq", "time")], as.character)

windows(width=19, height=10) 
par(las=1, cex.lab=0.75, cex.axis=0.6, bty="n", mgp=c(1, 0.6, 0), 
    oma=c(2, 4, 2, 0) + 0.1, mar=c(4, 0, 3, 3) + 0.1)
a <- layout(matrix(1:21, 3, 7, byrow=TRUE))   # with byrow
layout.show(a)

l <- split(df1, df1[c("freq", "time")])
m <- t(sapply(l, function(x) x[1, c(1, 3)]))  # matrix of first rows of each subset
m[, 2] <- sub("m", "", m[, 2])                # use the values...
m <- apply(m, 1:2, as.numeric)                # ... make numeric

Now we obtain the histograms within a lapply over the list ordered by m.

lapply(l[order(m[, 2], m[, 1])], function(x)
  hist(x$thr, main=paste("freq:", paste(x[1, c(1, 3)], collapse=","))))

New Result

Data

df1 <- structure(list(freq = structure(c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 
1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 
3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 
5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 
7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 
2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 
4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 
6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 
1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 
3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 
5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 
7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 
2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 
4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 
6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 
1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 
3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 
5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 
7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 
2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L, 2L, 3L, 
4L, 5L, 6L, 7L, 1L, 2L, 3L, 4L, 5L, 6L, 7L), .Label = c("4", 
"8", "12.5", "16", "20", "25", "31.5"), class = "factor"), thr = c(60L, 
25L, 20L, 15L, 15L, 30L, 35L, 60L, 25L, 10L, 15L, 15L, 30L, 35L, 
55L, 30L, 15L, 15L, 10L, 25L, 40L, 50L, 25L, 15L, 10L, 15L, 20L, 
40L, 50L, 30L, 10L, 15L, 15L, 20L, 25L, 50L, 25L, 10L, 10L, 10L, 
20L, 25L, 45L, 20L, 10L, 10L, 10L, 20L, 25L, 45L, 15L, 10L, 10L, 
10L, 20L, 30L, 60L, 30L, 10L, 10L, 10L, 15L, 30L, 50L, 25L, 10L, 
10L, 10L, 20L, 30L, 45L, 25L, 15L, 10L, 15L, 30L, 35L, 50L, 25L, 
15L, 10L, 15L, 25L, 35L, 60L, 25L, 10L, 10L, 15L, 20L, 30L, 60L, 
25L, 5L, 5L, 10L, 20L, 30L, 45L, 20L, 5L, 10L, 10L, 20L, 30L, 
45L, 20L, 10L, 10L, 10L, 20L, 30L, 60L, 30L, 15L, 10L, 15L, 25L, 
30L, 55L, 25L, 10L, 10L, 10L, 20L, 30L, 55L, 35L, 10L, 10L, 10L, 
20L, 30L, 60L, 35L, 15L, 10L, 10L, 15L, 25L, 50L, 30L, 10L, 10L, 
10L, 20L, 25L, 55L, 25L, 10L, 10L, 15L, 25L, 25L, 65L, 30L, 10L, 
10L, 15L, 20L, 30L, 60L, 30L, 15L, 15L, 15L, 15L, 30L, 55L, 35L, 
15L, 15L, 15L, 25L, 35L, 55L, 35L, 15L, 15L, 15L, 25L, 35L, 60L, 
35L, 15L, 15L, 15L, 25L, 35L, 60L, 30L, 10L, 10L, 15L, 25L, 35L, 
55L, 30L, 15L, 10L, 10L, 25L, 30L, 50L, 25L, 10L, 10L, 10L, 20L, 
30L, 55L, 30L, 10L, 10L, 15L, 20L, 30L, 55L, 30L, 10L, 15L, 20L, 
25L, 35L, 55L, 25L, 15L, 15L, 15L, 25L, 40L, 50L, 20L, 10L, 10L, 
20L, 30L, 40L, 45L, 25L, 10L, 10L, 10L, 20L, 30L, 50L, 25L, 10L, 
10L, 10L, 20L, 25L, 55L, 20L, 10L, 10L, 15L, 25L, 35L, 50L, 20L, 
10L, 10L, 15L, 25L, 30L, 45L, 20L, 15L, 10L, 10L, 20L, 30L, 50L, 
20L, 15L, 15L, 15L, 20L, 30L, 60L, 35L, 15L, 10L, 15L, 25L, 30L, 
60L, 35L, 15L, 15L, 15L, 30L, 35L, 55L, 25L, 10L, 15L, 15L, 25L, 
35L, 50L, 30L, 10L, 15L, 15L, 25L, 35L, 55L, 25L, 20L, 15L, 15L, 
25L, 30L, 55L, 25L, 15L, 15L, 15L, 30L, 35L), time = structure(c(1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 
3L), .Label = c("3m", "6m", "9m"), class = "factor")), row.names = c(NA, 
-322L), class = "data.frame")