R: Experiment 1: Uplink inter-arrival times stats
| library(plotrix);library(moments)foo<-function(m,n){v <- 1:12 | foo(From_Chip_RTP_delay, 10^5) | What to put into a report |
|---|---|---|
| v[1]=mean(m[1:n]) | "0.02000027577" | 0.020000 s |
| v[2]=std.error(m[1:n]) | "3.63331229733734e-07" | 3.63e-07 s |
| v[3]=names(sort(-table(m[1:n])))[1] | "0.0200049999984913" | 0.020005 s |
| v[4]=sd(m[1:n]) | "0.000114895423102849" | 0.000115 s |
| v[5]=var(m[1:n]) | "1.32009582499827e-08" | 1.32e-08 s |
| v[6]=kurtosis(m[1:n]) | "742.581556664333" | 742.58 |
| v[7]=skewness(m[1:n]) | "0.633658007213615" | 0.634 |
| v[8]=min(m[1:n]) | "0.0136249999995925" | 0.013625 s |
| v[9]=max(m[1:n]) | "0.0263840000006894" | 0.026384 s |
| v[10]=sum(m[1:n]) | "2000.027577" | 2000.027577 s |
| v[11]=length(m[1:n]) | "100000" | 100000 |
| v[12]=qnorm(0.965)*std.error(m[1:n]) | "6.58323732971544e-07" | 6.58e-07 s |
| return(v)} |
Truncated to a meaningful number of digits, added units, decimal align the numbers, set in a fixed-width font (Courier) |
Transcript
Now, one of the things that you're going to need to do, of courses, as you get out of either excel or whatever statistics package you get these huge big numbers, those are not useful. You need to figure out which of these digits are meaningful, and that's what you're going to put in the report. So, in this case, I'm going to say the mean value is 0.020000 seconds. Why? Because I know that I captured by data and the resolution of my clock is one microsecond - so there is no point in reporting something that has a higher resolution than what my clock actually had when I was sampling the data. So these extra digits over here aren't meaningful. And similarly, doing it for the other values. You note that decimal aligned them, which makes it much easier to be able to compare them, and I set them in a fixed-width font, which again makes it very, very easy to compare these numbers as I go down the column.