Graphics Reference
In-Depth Information
Figure 8-27. Left: scatter plot with log
10
x- and y-axes, and with manually specified breaks; right:
with exponents for the tick labels
To instead use exponential notation for the break labels (
Figure 8-27
, right), use the
trans_format()
function, from the scales package:
library(scales)
p
+
scale_x_log10(breaks
=
10
^
(
-1
:
5
),
labels
=
trans_format(
"log10"
, math_format(
10
^
.
x)))
+
scale_y_log10(breaks
=
10
^
(
0
:
3
),
labels
=
trans_format(
"log10"
, math_format(
10
^
.
x)))
Another way to use log axes is to transform the data before mapping it to the xand ycoordinates
(
Figure 8-28
)
. Technically, the axes are still linear—it's the quantity that is log-transformed:
ggplot(Animals, aes(x
=
log10(body), y
=
log10(brain), label
=
rownames(Animals)))
+
geom_text(size
=
3
)
The previous examples used a log
10
transformation, but it is possible to use other transforma-
these—
scale_x_log10()
is shorthand, but for these other log scales, we need to spell them
out:
library(scales)
# Use natural log on x, and log2 on y
p
+
scale_x_continuous(trans
=
log_trans(),
