我有以下数据框
Condition congruency Distance Measure Score liminf limsup ConditionF MeasureF
1 Zero Neutral 0 RTs 445.000 435.000 455.000 Zero RTs
2 Zero Congruent 1 RTs 445.000 435.000 456.000 Zero RTs
3 Zero Congruent 2 RTs 441.000 430.000 451.000 Zero RTs
4 Zero Congruent 3 RTs 432.000 422.000 442.000 Zero RTs
5 Zero Incongruent 1 RTs 449.000 439.000 459.000 Zero RTs
6 Zero Incongruent 2 RTs 449.000 438.000 460.000 Zero RTs
7 Zero Incongruent 3 RTs 453.000 440.000 465.000 Zero RTs
8 All different Neutral 0 RTs 446.000 436.000 456.000 All different RTs
9 All different Congruent 1 RTs 445.000 434.000 455.000 All different RTs
10 All different Congruent 2 RTs 449.000 438.000 461.000 All different RTs
11 All different Congruent 3 RTs 449.000 438.000 461.000 All different RTs
12 All different Incongruent 1 RTs 446.000 436.000 456.000 All different RTs
13 All different Incongruent 2 RTs 447.000 436.000 458.000 All different RTs
14 All different Incongruent 3 RTs 450.000 440.000 461.000 All different RTs
15 Zero Neutral 0 Errors 0.029 0.018 0.039 Zero Errors
16 Zero Congruent 1 Errors 0.023 0.015 0.031 Zero Errors
17 Zero Congruent 2 Errors 0.023 0.014 0.033 Zero Errors
18 Zero Congruent 3 Errors 0.027 0.018 0.036 Zero Errors
19 Zero Incongruent 1 Errors 0.034 0.024 0.044 Zero Errors
20 Zero Incongruent 2 Errors 0.036 0.024 0.048 Zero Errors
21 Zero Incongruent 3 Errors 0.024 0.013 0.035 Zero Errors
22 All different Neutral 0 Errors 0.019 0.013 0.026 All different Errors
23 All different Congruent 1 Errors 0.028 0.017 0.038 All different Errors
24 All different Congruent 2 Errors 0.021 0.011 0.032 All different Errors
25 All different Congruent 3 Errors 0.031 0.022 0.041 All different Errors
26 All different Incongruent 1 Errors 0.024 0.010 0.037 All different Errors
27 All different Incongruent 2 Errors 0.020 0.011 0.029 All different Errors
28 All different Incongruent 3 Errors 0.019 0.010 0.028 All different Errors
我想为RT和错误制作一个折线图（我的DV。因为我想将它们放在同一个图形空间中，我使用face_grid：
ggplot(matriz, aes(Distance, Score, shape= congruency, linetype=congruency)) +
geom_point(size=5) +
geom_line(size=1) +
geom_errorbar(aes(ymax = limsup, ymin= liminf), width=0.25, linetype=1) +
facet_grid(MeasureF~ConditionF,scales= "free")
问题是我想自定义每个y轴。具体来说，我想为此轴设置限制。如果我有一个简单的图形（我只需要包含指令coord_cartesian()），这很简单，但是当我使用facet_grid()时，我不知道如何做到这一点。
答案 0 :(得分：1)
如果要在ggplot2中扩展具有分面的绘图的y限制，可以执行以下方法：
library(ggplot2)
library(reshape2)
set.seed(1)
N <- 24
dates = seq(as.Date("2014-01-01"), as.Date("2015-12-01"), by = "1 month")
dummy_data <- data.frame(dates = dates, x = rnorm(N,0,1), group = sample(c("A", "B", "C"), size = N, replace = TRUE))
dummy_data_m <- melt(dummy_data, id.vars = c("dates", "group"))
plot = ggplot(data = dummy_data_m, aes(x = dates, y = value, colour = variable)) + geom_point() + facet_grid(group ~., scales="free_y")
print(plot) # Want to modify y-axis on this plot for each panel
# to expand plot axes, build the y ranges in a data.frame and add a new layer to the plot using geom_blank
# pick an x.value in the range of your x-values in dummy_data
x.value = as.Date("2014-01-01")
x.value = as.Date("2014-01-01")
# Say you want to expand the y-axis ranges for the different subpanels to be (-5, 5), (-4, 4), (-2, 2).
# If you simply plot at this point the y limits are roughly ~(-1.5, 1.5) for each plot
lower_y = data.frame(dates = x.value, group = c("A", "B", "C"), value = c(5, 4, 2))
y_ranges = rbind(lower_y, upper_y)
y_ranges_m = melt(y_ranges, id.vars = c("dates", "group"))
plot = plot + geom_blank(data = y_ranges_m, aes(x = dates, y = value, colour = variable))
print(plot)
对于相反的情况，如果要减小y范围，则从要缩小y轴的面板中的数据中删除观察值。例如（而不是如上所述扩展y轴）：
# do not want to plot any values where y < 0 for group A
dummy_data2 = dummy_data
x_adjusted = dummy_data2$x
x_adjusted[x_adjusted < 0 & dummy_data2$group == "A"] = NA
dummy_data2$x = x_adjusted
dummy_data_m2 <- melt(dummy_data2, id.vars = c("dates", "group"))
plot = ggplot(data = dummy_data_m2, aes(x = dates, y = value, colour = variable)) + geom_point() + facet_grid(group ~., scales="free_y")
print(plot) # this will show a plot where the panel for A has a y-axis lower limit above 0