
Exercise Title:
Creating a Frequency Distribution Graph (Histogram) in Excel for Integrated Data
Viewer (IDV) Grids

Abstract: Until
IDV gets its own histogram tool, this method works very quickly and
easily, and can be run alongside IDV with no trouble.

Preliminary Reading (in
OceanTeacher, unless otherwise indicated):

Required Software:

Other Resources:

Author:
Murray Brown

Version:
982014 (rerevised after much coffee)

1. This is the Excel spreadsheet you can export from IDV for any
raster product. Note that it is essentially an extended XYZ table,
with separate columns for each time step.

2. We can't work with the
#NUM! values, so use Excel's FIND AND REPLACE function to replace every
occurrence with nothing, i.e. delete them all. 

3. This message indicates
successful deletion. The table is ready for use. 

4. This is the spreadsheet
with the nodata values removed. We want to make a frequency distribution
(i.e. a histogram) for any of the data columns. We don't have time to
study them to find out the value ranges, so we need a quick trick to see
them automatically. 

5. Find the DATA ANALYSIS
control. If you can't find it, them you must install the ADD IN. 

6. Select the HISTOGRAM tool,
and click OK. 

7. For this example, let's
make a histogram for the E column of data values. Select the INPUT
RANGE control (the tiny spreadsheet icon to its right), and then select the
E column. Excel will automatically show $E:$E to indicate this
column. Then check these:
Then don't do anything else, but click OK. 

8. Here you see a "quick and
dirty" histogram for the whole dataset, made automatically by Excel.
You can easily see that the value range should be 3336 (or similar). 

9. Now, in the
real world, you'll need some value ranges, arranged into evenly spaced
"bins" for your data. Each type of data, e.g. salinity, temperature,
will have at least one favorite value range, based on the area where you
work, the seasons, etc. We provide you a generic salinity value range
set in he next panel, where you can copy and use the values you need to make
a really good histogram in Excel. The author has performed some
experiments with Excel, and discovered how the bin values work with the
HISTOGRAM data tool. You'll need to do work on your own later to
derive your own bin value scales. 
10. How distribution bin
values (in bold) work with the data value ("x") distributions in Excel:
 0: x = 0
 1: 0 < x <= 1, centered on 0.5
 2: 1 < x <= 2, centered on 1.5
 3: 2 < x <= 3, centered on 2.5
 etc.
So this means that the bin value is the maximum value of the data
in that bin. The minimum value of data in that bin depends on
the next lower bin value. 
11.
There is a very easy way to create a good set of bin values in Excel, which
we'll use below. But you must take responsibility for the selection of
the values and for defending your selections scientifically. 
12. Decide on the bin values
you want to use for your final histogram, based on the "quick and dirty" histogram.
For example, in the above example, we might chose these values:
 32
 32.5
 33
 33.5
 34
 34.5
 35
 35.5
 36
 36.5
 37

13. Make a new column anywhere
in your data spreadsheet and type the first few bin values into it (starting
with the lowest). 

14. Now Excel can finish the
list for you. Grab the tiny handle at the bottom right corner of the
selected area and slowly drag it down. Check the values that Excel
automatically adds, until you reach the limit you chose (in this case, 37). 

15. Now that you have a set
of bin values, you can run the histogram tool again. But this time you
should specify the column for the BIN RANGE. Excel "knows" exactly
what you're doing, and is quite flexible in how you make this designation.
Click OK to make the analysis, and be prepared to wait a few minutes for all
the calculations to finish.


16. This results in a nice
clean histogram with easily understood abscissa values. NOTE: The
graphic is extremely easy to edit, as you see here. 

17. Look at the bottom of the
Excel spreadsheet to see the tab structure. SHEET0 is the original
dataset, with the bin value column added. The other SHEETS are
different experiments you have performed, so you can go back and see
everything again. As usual, you can save the whole set of spreadsheets as
a workbook in Excel. 

18. You can set
up the workbook however you like for quickly adding data and making the
histograms. And you'll probably want to make a "canonical" set of bin
values for use with commonly displayed parameters. 