NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
> Hi Henner, > > > i have two functions: > > a) bucket -> standard deviation > > b) index -> (bucket, deviation) // where i mean by > > deviation: observed value - average > > > > and i want the values of function a) to be plotted as a thick line over and > > under the corresponding bucket's 0-value > > and the sample points of function b) just as points in the same display. of > > course i want both to use the same scale on the y-axis. so that one can > > clearly see when the deviation of a sample point is bigger than the standard > > deviation. > > > > it seems that i can't use the same RealTypes for mappings to ShapeScale and > > YAxis, because that way the shapes of function a) are drawn at the > > corresponding y-value instead of 0. if i use different RealTypes instead > > there seems to be no way to scale the shapes of function a to match the > > scale of function b) (i mean if if standard deviation is 0.5 and the > > observed deviation is also 0.5 the point of function b) should be exactly on > > top of the shape). > > You can use ScalarMaps of different RealTypes YAxis, if you > call ScalarMap.setRange(low, hi) for both ScalarMaps with > identical low and hi values. > > Please let me know if this doesn't solve your problem. no, sorry, that doesn't solve my problem, because the range of a RealType mapped to ShapeScale only seems to affect the size of the shape. if the minimum of the observed deviations is (let's say) -1 and i'd set this also as minimum for the range concerning ShapeScale nearly all the shapes would have same maximum lengths, because they represent (always positive) standard deviations (for example a std dev of 0 should be represented by bar of length 0, but it would be longer that way, because 0 is "much" bigger then the minimum (see assumption above). i'll try to explain my problem a bit more understandably: - i have a sequence of sampled data from which i compute the average and the standard deviation. - i want to plot that standard deviation as a bar (a line from +sigma to -sigma cutting the x-axis vertically). lets say its 1. - i want to plot a single point x (perhaps a new point of the considered sequence) with the same scale. if x < 1 it should be plotted on the bar, if x = -1 it should be plotted exactly at the lower end and if x = 1 exactly at the upper end of the bar. > > Cheers, > Bill > ---------------------------------------------------------- > Bill Hibbard, SSEC, 1225 W. Dayton St., Madison, WI 53706 > hibbard@xxxxxxxxxxxxxxxxx 608-263-4427 fax: 608-263-6738 > http://www.ssec.wisc.edu/~billh/vis.html thanks again henner
visad archives: