Outlier
An outlier is a value that sits far from the rest, like a $9,800 order among $300 ones. It can be a data-entry error or a real, important case: before you drop it, find out which.
Select any words here to ask your Maestro about them.
Into these 40 orders comes one for $9,800. Which number do you think moves more?
Why it matters to spot them: a single outlier can drag an average, inflate a total or warp a chart. Miss it and you report a number that describes almost nobody; delete it without asking and you may throw out the most important case of all (your biggest customer, the fraud, the day the system broke).
How to find them: plot the data. In a histogram they show up as stray bars far from the rest; in a box plot, as dots past the whiskers (the usual rule flags anything more than 1.5 times the box's height above or below the box). Sorting the column from largest to smallest works too.
What to do with one: find out whether it's an error (an extra zero, the wrong unit, a test row nobody deleted) or a real case. If it's an error, fix it at the source. If it's real, keep it and use measures that hold up, like the median, or report it on its own. Either way, write down what you decided.
You practice it in Is an extreme value an error or real data?.
How it's used
Before you drop it
The $9,800 order
- Is it possible? A store that sells items up to $2,000 can get an order like that.If it were impossible (an age of 240), it's a data-entry error.
- Are there others like it? Look for other orders over $5,000 this year.If there are several, it's a kind of customer, not a fluke.
- What does the source say? Open the order in the system or ask whoever entered it.
- marked: Decide and write it down: keep it, fix it or report it separately, and note why.What you don't do is delete it quietly.
When to look closer
Error or real case?
Five values far from the rest
- A customer's age: 240
- A $9,800 order among $300 ones
- A $0.01 order
- An order delivered in 0 days
- 113 °F in Phoenix on a July day
All five sit far from the rest. Which do you think are data-entry errors?
Ask your Maestro about Outlier
It answers from the free course page closest to this word, the way it answers while you study.
You: What does "Outlier" mean, in plain words?
Your Maestro: I'll explain it with a case of yours: it sticks better than a definition. What would you do with the $9,800 order if it were your store's?
Asking about: Answers from the page "Is an extreme value an error or real data?"
Only this browser keeps what you ask here. How it works
Learn more
Related terms
- Histogram
A histogram shows how the values in a column of numbers are spread out: it cuts them into ranges of equal width and draws one bar per range, as tall as the number of values that fall in it.
- Box plot
A box plot sums up a variable in five numbers: the minimum, the first quartile, the median, the third quartile and the maximum. The box runs from the first to the third quartile and holds the middle half of the data.
- Spot check
Taking a few rows spread across the data (from the start, the middle and the end) and comparing them by hand with what you expect.
- Edge case
An input at the edge of what the logic expects, like a code that isn't in the lookup table, to see that it fails the way it should instead of returning a plausible wrong value.