
34 Applying Analytics: A Practical Introduction
can be quite difficult because the mean hides many of the da ta set’s features.
Outliers can dominate
the mean, meaning they can exert disproportionately
strong influence on the algorithm’s output. Let’s illustrate the influence of
outliers on the mean and the concept of dominance with a notional example.
Imagine a sample of 50 people, 49 of which have an income of $50,000 each and
1 who has an income of $100 million; the average income of the sample is $2
million. Is this really a meaningful typical value? The income of the 49 people
could have been $0 and it would have barely changed the mean, be c ause