The exact number of bins you use isn't
critical but if you use too many you will
have too few values in each and risk missing
the big picture. On the other hand you
use too few you might miss important details.
There are various ways of calculating
the optimum number of bins. I use the
square root of the number of data values
as a starting point.
In the example there are 25 data values
so that gives 5 bins. The smallest data
value is 1 and the largest is 96. These
will fit nicely on a scale stretching
from 0 to 100 and the five bins will each
have a span 20.
Suppose there were 50 values. The square
root of 50 is just over 7. Keeping the
scale of 0 to 100 would give a bin span
of:
This would give starting values of:
But bin spans of 15 would be nicer numbers:
The square root rule tends to underestimate
the number of bins that give the best
result. With this in mind a bin span of
10 might work well:
0 
10 
20 
30 
40 
50 
60 
70 
80 
90 
100 
It only takes a couple of minutes to
create bins using Excel so I would probably
try both and see which looked better.

The data show the wait times,
in minutes, for 50 admissions
into the casualty department of
a hospital:
24 
22 
24 
30 
24 
16 
18 
32 
27 
69 
26 
36 
41 
27 
43 
29 
26 
21 
39 
44 
25 
32 
30 
28 
26 
34 
21 
30 
30 
31 
32 
37 
64 
26 
68 
20 
32 
43 
31 
24 
20 
27 
30 
33 
39 
40 
22 
31 
29 
43 
spreadsheet
Draw a histogram and comment
on the results. What action would
you suggest?

