Summary Statistics Grouped by Category

Note

The nominal and ordinal array data types are not recommended. To represent ordered and unordered discrete, nonnumeric data, use the Categorical Arrays data type instead.

Summary Statistics Grouped by Category

This example shows how to compute summary statistics grouped by levels of a categorical variable. You can compute group summary statistics for a numeric array or a dataset array using grpstats.

Load sample data.

load hospital

The dataset array, hospital, has 7 variables (columns) and 100 observations (rows).

Compute summary statistics by category.

The variable Sex is a nominal array with two levels, Male and Female. Compute the minimum and maximum weights for each gender.

stats = grpstats(hospital,'Sex',{'min','max'},'DataVars','Weight')

stats = 
              Sex       GroupCount    min_Weight    max_Weight
    Female    Female    53            111           147       
    Male      Male      47            158           202       

The dataset array, stats, has observations corresponding to the levels of the variable Sex. The variable min_Weight contains the minimum weight for each group, and the variable max_Weight contains the maximum weight for each group.

Compute summary statistics by multiple categories.

The variable Smoker is a logical array with value 1 for smokers and value 0 for nonsmokers. Compute the minimum and maximum weights for each gender and smoking combination.

stats = grpstats(hospital,{'Sex','Smoker'},{'min','max'},...
                 'DataVars','Weight')

stats = 
                Sex       Smoker    GroupCount    min_Weight    max_Weight
    Female_0    Female    false     40            111           147       
    Female_1    Female    true      13            115           146       
    Male_0      Male      false     26            158           194       
    Male_1      Male      true      21            164           202       

The dataset array, stats, has an observation row for each combination of levels of Sex and Smoker in the original data.