Univariate outlier detection: the IQR rule

Outlier detection is an important step in your exploratory data analysis. Anomalous observations (also known as outliers), if not properly handled, can skew your analysis and produce misleading conclusions.

Box plots help visually identify potential outliers as they summarize the distribution of a numerical variable. A commonly accepted rule of thumb is that an outlier is any value below \(Q1 - 1.5\times IQR\) or above \(Q3 + 1.5\times IQR\), where \(Q1\) and \(Q3\) are the first and third quartiles, respectively, of the variable distribution and \(IQR=Q3-Q1\) is the interquartile range.

In this exercise, you will apply the IQR rule to spot outliers in car fuel consumption. The cars dataset is already loaded. The quantile() function can be used to calculate \(Q1\) and \(Q3\).

Question

Why is important to detect and treat outliers?

Possible Answers

Because they are not representative of the data distribution.
Because they can drastically bias/change the fit estimates and predictions.
Because they make exploratory data analysis more difficult.
Because they tend to group together.

Exercise

Univariate outlier detection: the IQR rule

Instructions 1/4

Question

Possible Answers