library(rgl)

options(rgl.printRglwidget = TRUE) #You may not need this.
mydata = read.csv('datacloud.csv')
mn = apply(mydata,2,mean)
cov(mydata)
(eig = eigen(cov(mydata)))
r = range(mydata,mn + 10*eig$vec)

plot3d(mydata,xlim=r,ylim=r,zlim=r)

segments3d(c(mn[1],mn[1]+10*eig$vec[1,1]), c(mn[1],mn[2]+10*eig$vec[2,1]), c(mn[3],mn[3]+10*eig$vec[3,1]),col="red",lwd=3)
segments3d(c(mn[1],mn[1]+10*eig$vec[1,2]), c(mn[1],mn[2]+10*eig$vec[2,2]), c(mn[3],mn[3]+10*eig$vec[3,2]),col="blue",lwd=3)
segments3d(c(mn[1],mn[1]+10*eig$vec[1,3]), c(mn[1],mn[2]+10*eig$vec[2,3]), c(mn[3],mn[3]+10*eig$vec[3,3]),col="green",lwd=3)