Dear List, I have a dataset of high variability. I conducted epoch analysis and attempted to plot the standard error bar alongside. I am, however, surprised that the error bars are of equal length. I do not think that the variability in the data is captured, except there is a kind of averaging that smooths out the differences in daily variations. Should that be the case, I don't know how and as such need your assistance to explain what is going on. The plot is attached. 71 events are represented in the plot. The code I use to generate the plot is: oodf<-data.frame(A,B) library(plotrix) std.error<-function(x) return(sd(x)/(sum(!is.na(x)))) oomean<-as.vector(by(oodf$B,oodf$A,mean)) oose<-as.vector(by(oodf$B,oodf$A,std.error)) plot(-5:10,oomean,type="b",ylim=c(145000,162000), xlab="days (epoch is the day of Fd)",ylab="strikes/day",main="Superposed Epoch of all the Events") dispersion(-5:10,oomean,oose). The sample data is: -5 64833 -4 95864 -3 82322 -2 95591 -1 69378 0 74281 1 103261 2 92473 3 84344 4 127415 5 123826 6 100029 7 76205 8 105162 9 119533 10 106490 -5 82322 -4 95591 -3 69378 -2 74281 -1 103261 0 92473 1 84344 2 127415 3 123826 4 100029 5 76205 6 105162 7 119533 8 106490 9 114771 10 55593 -5 85694 -4 65205 -3 80995 -2 51723 -1 62310 0 53401 1 65677 2 76094 3 64035 4 68290 5 73306 6 82176 7 75566 8 89762 9 88063 10 94395 -5 80651 -4 81291 -3 63702 -2 70297 -1 64117 0 71219 1 57354 2 62111 3 42252 4 35454 5 33469 6 38899 7 64981 8 85694 9 79452 10 85216 -5 71219 -4 57354 -3 62111 -2 42252 -1 35454 0 33469 1 38899 2 64981 3 85694 4 79452 5 85216 6 81721 7 91231 8 107074 9 108103 10 75768 You kind help will be greatly appreciated. Many thanks Ogbos -------------- next part -------------- A non-text attachment was scrubbed... Name: A8.png Type: image/png Size: 10229 bytes Desc: not available URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20180904/a9a19f5a/attachment-0002.png>
Thank you for the reproducible data, but it is not the data used in the plot you attached and does not plot anything with the code you included. The ylim= argument must be modified: plot(-5:10, oomean, type="b", ylim=c(40000, 120000), xlab="days (epoch is the day of Fd)", ylab="strikes/day", main="Superposed Epoch of all the Events") dispersion(-5:10, oomean, oose, arrow.cap=.01) On the plot of these data it is clear that the error bars are different sizes: range(oose) [1] 1728.234 6890.916 What was the range of oose values for the data in the plot you included with your message? ---------------------------------------- David L Carlson Department of Anthropology Texas A&M University College Station, TX 77843-4352 -----Original Message----- From: R-help <r-help-bounces at r-project.org> On Behalf Of Ogbos Okike Sent: Tuesday, September 4, 2018 2:35 PM To: r-help <r-help at r-project.org> Subject: [R] Equal Standard Error bar Dear List, I have a dataset of high variability. I conducted epoch analysis and attempted to plot the standard error bar alongside. I am, however, surprised that the error bars are of equal length. I do not think that the variability in the data is captured, except there is a kind of averaging that smooths out the differences in daily variations. Should that be the case, I don't know how and as such need your assistance to explain what is going on. The plot is attached. 71 events are represented in the plot. The code I use to generate the plot is: oodf<-data.frame(A,B) library(plotrix) std.error<-function(x) return(sd(x)/(sum(!is.na(x)))) oomean<-as.vector(by(oodf$B,oodf$A,mean)) oose<-as.vector(by(oodf$B,oodf$A,std.error)) plot(-5:10,oomean,type="b",ylim=c(145000,162000), xlab="days (epoch is the day of Fd)",ylab="strikes/day",main="Superposed Epoch of all the Events") dispersion(-5:10,oomean,oose). The sample data is: -5 64833 -4 95864 -3 82322 -2 95591 -1 69378 0 74281 1 103261 2 92473 3 84344 4 127415 5 123826 6 100029 7 76205 8 105162 9 119533 10 106490 -5 82322 -4 95591 -3 69378 -2 74281 -1 103261 0 92473 1 84344 2 127415 3 123826 4 100029 5 76205 6 105162 7 119533 8 106490 9 114771 10 55593 -5 85694 -4 65205 -3 80995 -2 51723 -1 62310 0 53401 1 65677 2 76094 3 64035 4 68290 5 73306 6 82176 7 75566 8 89762 9 88063 10 94395 -5 80651 -4 81291 -3 63702 -2 70297 -1 64117 0 71219 1 57354 2 62111 3 42252 4 35454 5 33469 6 38899 7 64981 8 85694 9 79452 10 85216 -5 71219 -4 57354 -3 62111 -2 42252 -1 35454 0 33469 1 38899 2 64981 3 85694 4 79452 5 85216 6 81721 7 91231 8 107074 9 108103 10 75768 You kind help will be greatly appreciated. Many thanks Ogbos -------------- next part -------------- A non-text attachment was scrubbed... Name: plot.png Type: image/png Size: 7334 bytes Desc: plot.png URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20180904/8f32584f/attachment-0002.png>
Hi David, You are right. Thanks for your time. The problem is certainly with the data. Plotting part of it gives different results, usually quite different from the output when the total data is used. In fact, just as you mentioned, it will look as if it is not the same code that is used to plot it. The range is 33469-281856. Instead of taking time to explain how I fiddled with the range before getting the plot I attached, let me just attach the whole data. I really had issues with the range when plotting the whole data (see attached please). Many thanks again. Ogbos On Tue, Sep 4, 2018 at 9:58 PM David L Carlson <dcarlson at tamu.edu> wrote:> Thank you for the reproducible data, but it is not the data used in the > plot you attached and does not plot anything with the code you included. > The ylim= argument must be modified: > > plot(-5:10, oomean, type="b", ylim=c(40000, 120000), > xlab="days (epoch is the day of Fd)", ylab="strikes/day", > main="Superposed Epoch of all the Events") > dispersion(-5:10, oomean, oose, arrow.cap=.01) > > On the plot of these data it is clear that the error bars are different > sizes: > > range(oose) > [1] 1728.234 6890.916 > > What was the range of oose values for the data in the plot you included > with your message? > > ---------------------------------------- > David L Carlson > Department of Anthropology > Texas A&M University > College Station, TX 77843-4352 > > -----Original Message----- > From: R-help <r-help-bounces at r-project.org> On Behalf Of Ogbos Okike > Sent: Tuesday, September 4, 2018 2:35 PM > To: r-help <r-help at r-project.org> > Subject: [R] Equal Standard Error bar > > Dear List, > > I have a dataset of high variability. I conducted epoch analysis and > attempted to plot the standard error bar alongside. > > I am, however, surprised that the error bars are of equal length. I do not > think that the variability in the data is captured, except there is a kind > of averaging that smooths out the differences in daily variations. Should > that be the case, I don't know how and as such need your assistance to > explain what is going on. > > The plot is attached. 71 events are represented in the plot. > > The code I use to generate the plot is: > oodf<-data.frame(A,B) > library(plotrix) > std.error<-function(x) return(sd(x)/(sum(!is.na(x)))) > oomean<-as.vector(by(oodf$B,oodf$A,mean)) > oose<-as.vector(by(oodf$B,oodf$A,std.error)) > plot(-5:10,oomean,type="b",ylim=c(145000,162000), > xlab="days (epoch is the day of Fd)",ylab="strikes/day",main="Superposed > Epoch of all the Events") > dispersion(-5:10,oomean,oose). > > The sample data is: > -5 64833 > -4 95864 > -3 82322 > -2 95591 > -1 69378 > 0 74281 > 1 103261 > 2 92473 > 3 84344 > 4 127415 > 5 123826 > 6 100029 > 7 76205 > 8 105162 > 9 119533 > 10 106490 > -5 82322 > -4 95591 > -3 69378 > -2 74281 > -1 103261 > 0 92473 > 1 84344 > 2 127415 > 3 123826 > 4 100029 > 5 76205 > 6 105162 > 7 119533 > 8 106490 > 9 114771 > 10 55593 > -5 85694 > -4 65205 > -3 80995 > -2 51723 > -1 62310 > 0 53401 > 1 65677 > 2 76094 > 3 64035 > 4 68290 > 5 73306 > 6 82176 > 7 75566 > 8 89762 > 9 88063 > 10 94395 > -5 80651 > -4 81291 > -3 63702 > -2 70297 > -1 64117 > 0 71219 > 1 57354 > 2 62111 > 3 42252 > 4 35454 > 5 33469 > 6 38899 > 7 64981 > 8 85694 > 9 79452 > 10 85216 > -5 71219 > -4 57354 > -3 62111 > -2 42252 > -1 35454 > 0 33469 > 1 38899 > 2 64981 > 3 85694 > 4 79452 > 5 85216 > 6 81721 > 7 91231 > 8 107074 > 9 108103 > 10 75768 > > > You kind help will be greatly appreciated. > > Many thanks > > Ogbos >