将均值和 sem 叠加到每个方面的 ggplot
overlay mean and sem to ggplot per facet
我有一个可怕的线图，其中我使用小平面来划分组（受过训练和未受过训练）和试验（E1、E2）。
我试着看看趋势会是什么样子，所以我使用了 stat_smooth() 但这不是我希望得到的。请注意红色平滑线代表数据的程度（特别是在由垂直线分隔的区域中）。
我现在想我更愿意以相同的方式绘制均值和标准误差。但我不知道该怎么做。我看过其他帖子，但我只找到了一个变量的简单点图的信息，没有任何重复。该图来自两组中每组的 20 个重复，采样率为 0.2 秒，因此它更复杂。
我是否应该找到一种方法来计算每组的均值和标准误差并将其传递给 stat_smooth() 函数？
很遗憾，我的数据超出了字符数限制，所以我将提供一个玩具示例。
dput(toy)
structure(list(tiempo = c(618.2, 618.4, 618.6, 618.8, 619, 619.2,
619.4, 619.6, 619.8, 620, 620.2, 620.4, 620.6, 620.8, 621, 621.2,
621.4, 621.6, 621.8, 622, 622.2, 622.4, 622.6, 622.8, 623, 623.2,
623.4, 623.6, 623.8, 624, 624.2, 624.4, 624.6, 624.8, 625, 625.2,
625.4, 625.6, 625.8, 626, 626.2, 626.4, 626.6, 626.8, 627, 3138.2,
3138.4, 3138.6, 3138.8, 3139, 3139.2, 3139.4, 3139.6, 3139.8,
3140, 3140.2, 3140.4, 3140.6, 3140.8, 3141, 3141.2, 3141.4, 3141.6,
3141.8, 3142, 3142.2, 3142.4, 3142.6, 3142.8, 3143, 3143.2, 3143.4,
3143.6, 3143.8, 3144, 3144.2, 3144.4, 3144.6, 3144.8, 3145, 3145.2,
3145.4, 3145.6, 3145.8, 3146, 3146.2, 3146.4, 3146.6, 3146.8,
3147), UT2x45.19 = c(8325L, 8324L, 8323L, 8321L, 8320L, 8317L,
8315L, 8313L, 8311L, 8307L, 8305L, 8302L, 8301L, 8298L, 8295L,
8294L, 8293L, 8292L, 8290L, 8287L, 8285L, 8283L, 8282L, 8281L,
8279L, 8275L, 8271L, 8270L, 8266L, 8263L, 8259L, 8257L, 8255L,
8254L, 8253L, 8252L, 8250L, 8246L, 8239L, 8235L, 8235L, 8234L,
8232L, 8229L, 8226L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L), UT2x45.20 = c(17444L, 17445L, 17447L, 17448L,
17449L, 17449L, 17451L, 17451L, 17452L, 17451L, 17451L, 17451L,
17451L, 17452L, 17453L, 17459L, 17461L, 17463L, 17462L, 17462L,
17463L, 17466L, 17467L, 17467L, 17467L, 17466L, 17467L, 17467L,
17468L, 17469L, 17469L, 17470L, 17471L, 17473L, 17475L, 17475L,
17475L, 17475L, 17476L, 17477L, 17479L, 17481L, 17482L, 17482L,
17483L, 19867L, 19868L, 19869L, 19869L, 19869L, 19869L, 19869L,
19869L, 19869L, 19869L, 19869L, 19869L, 19869L, 19869L, 19870L,
19870L, 19870L, 19870L, 19870L, 19870L, 19871L, 19871L, 19872L,
19873L, 19875L, 19877L, 19878L, 19878L, 19878L, 19878L, 19879L,
19879L, 19879L, 19881L, 19882L, 19883L, 19884L, 19884L, 19887L,
19887L, 19888L, 19890L, 19892L, 19892L, 19892L), TR2x45.19 = c(15870L,
15870L, 15866L, 15860L, 15861L, 15863L, 15863L, 15872L, 15877L,
15857L, 15845L, 15842L, 15840L, 15830L, 15826L, 15824L, 15818L,
15815L, 15813L, 15810L, 15808L, 15804L, 15799L, 15798L, 15776L,
15746L, 15734L, 15727L, 15723L, 15723L, 15709L, 15671L, 15634L,
15617L, 15598L, 15582L, 15568L, 15557L, 15547L, 15537L, 15526L,
15517L, 15506L, 15501L, 15496L, 14338L, 14338L, 14338L, 14338L,
14338L, 14338L, 14338L, 14338L, 14337L, 14329L, 14324L, 14323L,
14322L, 14322L, 14322L, 14321L, 14321L, 14321L, 14321L, 14321L,
14321L, 14321L, 14321L, 14321L, 14321L, 14313L, 14309L, 14307L,
14306L, 14307L, 14304L, 14280L, 14253L, 14239L, 14234L, 14225L,
14218L, 14204L, 14193L, 14183L, 14174L, 14163L, 14153L, 14143L,
14138L), TR2x45.20 = c(15760L, 15760L, 15761L, 15768L, 15780L,
15781L, 15787L, 15795L, 15795L, 15791L, 15791L, 15791L, 15789L,
15790L, 15792L, 15795L, 15794L, 15794L, 15793L, 15793L, 15791L,
15791L, 15791L, 15793L, 15795L, 15798L, 15806L, 15809L, 15812L,
15818L, 15822L, 15825L, 15829L, 15833L, 15837L, 15838L, 15839L,
15840L, 15841L, 15843L, 15845L, 15847L, 15849L, 15851L, 15854L,
14945L, 14945L, 14945L, 14944L, 14942L, 14943L, 14943L, 14943L,
14943L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L,
14942L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L,
14942L, 14941L, 14939L, 14939L, 14939L, 14939L, 14939L, 14939L,
14938L, 14938L, 14937L, 14936L, 14936L, 14936L, 14936L, 14936L,
14936L, 14936L, 14935L, 14935L, 14935L), ensayo = 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, 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), .Label = c("E1", "E2"), class = "factor")), .Names = c("tiempo",
"UT2x45.19", "UT2x45.20", "TR2x45.19", "TR2x45.20", "ensayo"), row.names = c("3088",
"3089", "3090", "3091", "3092", "3093", "3094", "3095", "3096",
"3097", "3098", "3099", "3100", "3101", "3102", "3103", "3104",
"3105", "3106", "3107", "3108", "3109", "3110", "3111", "3112",
"3113", "3114", "3115", "3116", "3117", "3118", "3119", "3120",
"3121", "3122", "3123", "3124", "3125", "3126", "3127", "3128",
"3129", "3130", "3131", "3132", "15688", "15689", "15690", "15691",
"15692", "15693", "15694", "15695", "15696", "15697", "15698",
"15699", "15700", "15701", "15702", "15703", "15704", "15705",
"15706", "15707", "15708", "15709", "15710", "15711", "15712",
"15713", "15714", "15715", "15716", "15717", "15718", "15719",
"15720", "15721", "15722", "15723", "15724", "15725", "15726",
"15727", "15728", "15729", "15730", "15731", "15732"), class = "data.frame")
以及函数
deriv_abs<- function(tiempo,dataframe){
matriz = as.matrix(dataframe)
matriz_dif = abs(diff(matriz ))# derivar
matriz_dif1= rbind(c(rep(0,dim(matriz_dif)[2])),matriz_dif) #agrega linea de ceros para compensar diff
return(data.frame(cbind(tiempo, matriz_dif1)))
}
vel_us<-by( toy[,(2:length(toy))-1],
toy$ensayo,
FUN= function(x) deriv_abs(x$tiempo, x[,-1] ))
names(vel_us$E1)<-names(toy[2:length(toy)-1])
names(vel_us$E2)<-names(toy[2:length(toy)-1])
df4<-data.frame(rbind(vel_us$E1,vel_us$E2),ensayo=us$ensayo)
m_absvel_us<-melt(df4,id=c("ensayo","tiempo"))
m_absvel_us$trat<-rep(NA,length(m_absvel_us))
m_absvel_us$trat[grep("TR",m_absvel_us$variable)]<-c("TR")
m_absvel_us$trat[grep("UT",m_absvel_us$variable)]<-c("UT")
us_absvel_plot<-ggplot(m_absvel_us,aes(x=tiempo,y=value,fill=trat,color=trat))+theme_minimal()
us_absvel_plot+geom_line(aes(fill=variable))+facet_wrap(trat~ensayo ,scales="free_x")+labs(title=c("Velocidad por ensayo entrenamiento")) + geom_vline(aes(xintercept = tiempo),lty=2, vline.data) +stat_smooth(aes(fill=trat))
有关更多详细信息，我添加了刻面第一行的图（数据的平均值）。这就是我想要覆盖的内容（或者可能是包含 SEM 的平滑版本）
要添加均值线和均值的标准误差：
us_absvel_plot +
stat_summary(fun.y=mean, geom="line", colour="red") +
stat_summary(fun.data=mean_cl_normal, geom="errorbar", confint=0.683, colour="red") +
stat_summary(fun.y=mean, geom="point", colour="blue")
将 conf.int 设置为 0.683 可以得到等于 +/- 1 个标准误差的误差条。
要更平滑地跟踪数据，有多种选择。例如，如果您使用 "loess" 平滑器（每组观测值少于 1000 的数据的默认值），您可以这样做：
geom_smooth(method="loess", span=1, se=FALSE, colour="red", size=2)
您将不得不使用 span 的值。跨度越小，平滑函数越灵活。
您也可以试试自然样条。在这种情况下，您设置更高的自由度，df，以获得更灵活的功能：
library(splines)
geom_smooth(method="lm", formula=y ~ ns(x,df=10), colour="red", size=2)
我有一个可怕的线图，其中我使用小平面来划分组（受过训练和未受过训练）和试验（E1、E2）。
我试着看看趋势会是什么样子，所以我使用了 stat_smooth() 但这不是我希望得到的。请注意红色平滑线代表数据的程度（特别是在由垂直线分隔的区域中）。
我现在想我更愿意以相同的方式绘制均值和标准误差。但我不知道该怎么做。我看过其他帖子，但我只找到了一个变量的简单点图的信息，没有任何重复。该图来自两组中每组的 20 个重复，采样率为 0.2 秒，因此它更复杂。
我是否应该找到一种方法来计算每组的均值和标准误差并将其传递给 stat_smooth() 函数？
很遗憾，我的数据超出了字符数限制，所以我将提供一个玩具示例。
dput(toy)
structure(list(tiempo = c(618.2, 618.4, 618.6, 618.8, 619, 619.2,
619.4, 619.6, 619.8, 620, 620.2, 620.4, 620.6, 620.8, 621, 621.2,
621.4, 621.6, 621.8, 622, 622.2, 622.4, 622.6, 622.8, 623, 623.2,
623.4, 623.6, 623.8, 624, 624.2, 624.4, 624.6, 624.8, 625, 625.2,
625.4, 625.6, 625.8, 626, 626.2, 626.4, 626.6, 626.8, 627, 3138.2,
3138.4, 3138.6, 3138.8, 3139, 3139.2, 3139.4, 3139.6, 3139.8,
3140, 3140.2, 3140.4, 3140.6, 3140.8, 3141, 3141.2, 3141.4, 3141.6,
3141.8, 3142, 3142.2, 3142.4, 3142.6, 3142.8, 3143, 3143.2, 3143.4,
3143.6, 3143.8, 3144, 3144.2, 3144.4, 3144.6, 3144.8, 3145, 3145.2,
3145.4, 3145.6, 3145.8, 3146, 3146.2, 3146.4, 3146.6, 3146.8,
3147), UT2x45.19 = c(8325L, 8324L, 8323L, 8321L, 8320L, 8317L,
8315L, 8313L, 8311L, 8307L, 8305L, 8302L, 8301L, 8298L, 8295L,
8294L, 8293L, 8292L, 8290L, 8287L, 8285L, 8283L, 8282L, 8281L,
8279L, 8275L, 8271L, 8270L, 8266L, 8263L, 8259L, 8257L, 8255L,
8254L, 8253L, 8252L, 8250L, 8246L, 8239L, 8235L, 8235L, 8234L,
8232L, 8229L, 8226L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L, 1579L,
1579L, 1579L, 1579L), UT2x45.20 = c(17444L, 17445L, 17447L, 17448L,
17449L, 17449L, 17451L, 17451L, 17452L, 17451L, 17451L, 17451L,
17451L, 17452L, 17453L, 17459L, 17461L, 17463L, 17462L, 17462L,
17463L, 17466L, 17467L, 17467L, 17467L, 17466L, 17467L, 17467L,
17468L, 17469L, 17469L, 17470L, 17471L, 17473L, 17475L, 17475L,
17475L, 17475L, 17476L, 17477L, 17479L, 17481L, 17482L, 17482L,
17483L, 19867L, 19868L, 19869L, 19869L, 19869L, 19869L, 19869L,
19869L, 19869L, 19869L, 19869L, 19869L, 19869L, 19869L, 19870L,
19870L, 19870L, 19870L, 19870L, 19870L, 19871L, 19871L, 19872L,
19873L, 19875L, 19877L, 19878L, 19878L, 19878L, 19878L, 19879L,
19879L, 19879L, 19881L, 19882L, 19883L, 19884L, 19884L, 19887L,
19887L, 19888L, 19890L, 19892L, 19892L, 19892L), TR2x45.19 = c(15870L,
15870L, 15866L, 15860L, 15861L, 15863L, 15863L, 15872L, 15877L,
15857L, 15845L, 15842L, 15840L, 15830L, 15826L, 15824L, 15818L,
15815L, 15813L, 15810L, 15808L, 15804L, 15799L, 15798L, 15776L,
15746L, 15734L, 15727L, 15723L, 15723L, 15709L, 15671L, 15634L,
15617L, 15598L, 15582L, 15568L, 15557L, 15547L, 15537L, 15526L,
15517L, 15506L, 15501L, 15496L, 14338L, 14338L, 14338L, 14338L,
14338L, 14338L, 14338L, 14338L, 14337L, 14329L, 14324L, 14323L,
14322L, 14322L, 14322L, 14321L, 14321L, 14321L, 14321L, 14321L,
14321L, 14321L, 14321L, 14321L, 14321L, 14313L, 14309L, 14307L,
14306L, 14307L, 14304L, 14280L, 14253L, 14239L, 14234L, 14225L,
14218L, 14204L, 14193L, 14183L, 14174L, 14163L, 14153L, 14143L,
14138L), TR2x45.20 = c(15760L, 15760L, 15761L, 15768L, 15780L,
15781L, 15787L, 15795L, 15795L, 15791L, 15791L, 15791L, 15789L,
15790L, 15792L, 15795L, 15794L, 15794L, 15793L, 15793L, 15791L,
15791L, 15791L, 15793L, 15795L, 15798L, 15806L, 15809L, 15812L,
15818L, 15822L, 15825L, 15829L, 15833L, 15837L, 15838L, 15839L,
15840L, 15841L, 15843L, 15845L, 15847L, 15849L, 15851L, 15854L,
14945L, 14945L, 14945L, 14944L, 14942L, 14943L, 14943L, 14943L,
14943L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L,
14942L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L, 14942L,
14942L, 14941L, 14939L, 14939L, 14939L, 14939L, 14939L, 14939L,
14938L, 14938L, 14937L, 14936L, 14936L, 14936L, 14936L, 14936L,
14936L, 14936L, 14935L, 14935L, 14935L), ensayo = 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, 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), .Label = c("E1", "E2"), class = "factor")), .Names = c("tiempo",
"UT2x45.19", "UT2x45.20", "TR2x45.19", "TR2x45.20", "ensayo"), row.names = c("3088",
"3089", "3090", "3091", "3092", "3093", "3094", "3095", "3096",
"3097", "3098", "3099", "3100", "3101", "3102", "3103", "3104",
"3105", "3106", "3107", "3108", "3109", "3110", "3111", "3112",
"3113", "3114", "3115", "3116", "3117", "3118", "3119", "3120",
"3121", "3122", "3123", "3124", "3125", "3126", "3127", "3128",
"3129", "3130", "3131", "3132", "15688", "15689", "15690", "15691",
"15692", "15693", "15694", "15695", "15696", "15697", "15698",
"15699", "15700", "15701", "15702", "15703", "15704", "15705",
"15706", "15707", "15708", "15709", "15710", "15711", "15712",
"15713", "15714", "15715", "15716", "15717", "15718", "15719",
"15720", "15721", "15722", "15723", "15724", "15725", "15726",
"15727", "15728", "15729", "15730", "15731", "15732"), class = "data.frame")
以及函数
deriv_abs<- function(tiempo,dataframe){
matriz = as.matrix(dataframe)
matriz_dif = abs(diff(matriz ))# derivar
matriz_dif1= rbind(c(rep(0,dim(matriz_dif)[2])),matriz_dif) #agrega linea de ceros para compensar diff
return(data.frame(cbind(tiempo, matriz_dif1)))
}
vel_us<-by( toy[,(2:length(toy))-1],
toy$ensayo,
FUN= function(x) deriv_abs(x$tiempo, x[,-1] ))
names(vel_us$E1)<-names(toy[2:length(toy)-1])
names(vel_us$E2)<-names(toy[2:length(toy)-1])
df4<-data.frame(rbind(vel_us$E1,vel_us$E2),ensayo=us$ensayo)
m_absvel_us<-melt(df4,id=c("ensayo","tiempo"))
m_absvel_us$trat<-rep(NA,length(m_absvel_us))
m_absvel_us$trat[grep("TR",m_absvel_us$variable)]<-c("TR")
m_absvel_us$trat[grep("UT",m_absvel_us$variable)]<-c("UT")
us_absvel_plot<-ggplot(m_absvel_us,aes(x=tiempo,y=value,fill=trat,color=trat))+theme_minimal()
us_absvel_plot+geom_line(aes(fill=variable))+facet_wrap(trat~ensayo ,scales="free_x")+labs(title=c("Velocidad por ensayo entrenamiento")) + geom_vline(aes(xintercept = tiempo),lty=2, vline.data) +stat_smooth(aes(fill=trat))
有关更多详细信息，我添加了刻面第一行的图（数据的平均值）。这就是我想要覆盖的内容（或者可能是包含 SEM 的平滑版本）
要添加均值线和均值的标准误差：
us_absvel_plot +
stat_summary(fun.y=mean, geom="line", colour="red") +
stat_summary(fun.data=mean_cl_normal, geom="errorbar", confint=0.683, colour="red") +
stat_summary(fun.y=mean, geom="point", colour="blue")
将 conf.int 设置为 0.683 可以得到等于 +/- 1 个标准误差的误差条。
要更平滑地跟踪数据，有多种选择。例如，如果您使用 "loess" 平滑器（每组观测值少于 1000 的数据的默认值），您可以这样做：
geom_smooth(method="loess", span=1, se=FALSE, colour="red", size=2)
您将不得不使用 span 的值。跨度越小，平滑函数越灵活。
您也可以试试自然样条。在这种情况下，您设置更高的自由度，df，以获得更灵活的功能：
library(splines)
geom_smooth(method="lm", formula=y ~ ns(x,df=10), colour="red", size=2)