Reply to the post by @kochavak:

1. Count number of members of each group (may use pivot table to count in multiple dimensions (i.e. may be 30 women but only 10 women under age 45)
2. Use ranges or other groupings. Age groups, larger geographic notations
3. Use quartile or percentiles (QUARTILE, PERCENTILE.INC formulas in Excel ) to define reasonable ranges. At any data set exactly 1/4 of the observations in each quartile, so see what the quartiles values are and refine readable ranges nerby if you want 4 groups or see percentiles to fine-tuning ( define your number of equal groups then the whole is 100).

Looks like what you described is a form of achieving k-anonymity.
Concepts extending that are l-diversity and t-closeness.