Detecting outliers

- [Instructor] Data scientists often run into data sets…that contain values that are well outside of the norm.…These unusual values can present special challenges…for data analysis, and it's important to understand…what they mean when they're present in a data set.…Outliers are data points that lie far outside the norm,…and they may occur for two reasons.…First, outliers may indicate…some type of error in the data set.…Someone may have measured data…incorrectly in the first place,…incorrectly input into a system,…or performed a calculation improperly.…

For example, imagine that you're looking at a data set…consisting of temperatures from…weather stations in New York,…and you find that there is a single data point…recording a temperature of 212 degrees Fahrenheit.…This is clearly some type of error.…Perhaps the thermometer failed.…Maybe somebody wrote down the temperature incorrectly,…or it could be that the thermometer…was misplaced inside of an oven.…Whatever the cause, this is clearly an invalid data point…for outdoor temperatures.…