Le vendredi 12 septembre 2008 à 15:28 -0400, Marlon Nelson a écrit : > Thank you, Jean > > Another question: The data table was produced using the Histogram > tool. The 1% bin counts 1011 data points between -1% and 1%. In the > chart of the histogram, the area representing those points falls > between 1% and 3% on the x-axis. Seems like an off-by-one problem to > me.
The histogram plot needs one more x data than y data the first y data is assumed to represent the number of occurences between the two first x data. I have plans to implement hidtograms from raw data at some moment in the future. Regards, Jean > On Fri, Sep 12, 2008 at 11:17 AM, Jean Bréfort <[EMAIL PROTECTED]> wrote: > > Le vendredi 12 septembre 2008 à 10:46 -0400, Marlon Nelson a écrit : > >> I originally thought this was a bug and was about 30 seconds from > >> submitting one to bugzilla when I began to suspect the problem is with > >> my understanding of what a histogram is. > >> > >> With the data listed below (frequency of daily returns of Merrill > >> Lynch stock over the last 10 years), I created a histogram chart. The > >> highest point on the chart reaches 50,000. I was expecting 1011. > >> > >> Reading a bit from wikipedia, I see what I was actually expecting to > >> see is a bar chart. > >> > >> But given a histogram chart of this data, what do the y-axis numbers mean? > >> > >> Bin Frequency > >> -15% 1 > >> -13% 1 > >> -11% 4 > >> -9% 6 > >> -7% 13 > >> -5% 53 > >> -3% 167 > >> -1% 510 > >> 1% 1011 > >> 3% 489 > >> 5% 156 > >> 7% 57 > >> 9% 26 > >> 11% 6 > >> 13% 4 > >> 15% 3 > >> 17% 2 > >> > >> -- > >> -eom- > > > > The histogram plots the density, as the 1011 data are in a 0.02 > > interval, you get 1011 / 0.02 = 50550 as the largest value. > > > > Regards, > > Jean > > > > > > > _______________________________________________ gnumeric-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/gnumeric-list
