The aggregation that is assigned to a measure determines how its values are summarized
in a
visualization or report object.
For example, in a
bar chart of Sales by Quarter, each bar represents the aggregated values of the Sales measure
for a specific quarter. If the aggregation for Sales is
Sum, then the bars represent the
sum (total) of sales for each quarter. If the aggregation for Sales
is
Average, then the bars represent the average
sales for each quarter.
Note: Some aggregation types can
override the data format that is used to display values in a visualization
or report object. For example, if a measure has the Currency format
with zero decimal places of precision, and you apply the Variance aggregation,
then the values are displayed using the Comma format with two decimal
places of precision instead.
You can specify the
following aggregations for your measures:
Sum
calculates the sum
(total) of the values of a measure.
Average
calculates the average
(mean) value of a measure.
Variance
calculates the variance
of a measure.
Minimum
calculates the smallest
value of a measure.
calculates the first
quartile of a measure.
calculates the third quartile of a measure.
Maximum
calculates the largest
value of a measure.
Skewness
calculates the
skewness of a measure. Skewness indicates the distribution of values. A positive value indicates
that the distribution is heavier for values greater than the mean. A negative value
indicates that the distribution is heavier for values less than the mean.
Kurtosis
calculates the kurtosis
of a measure. The kurtosis value indicates how peaked the distribution
is. A larger value indicates a more sharply peaked distribution. A
smaller value indicates a flatter distribution.
calculates the
coefficient of variation of a measure. The coefficient of variation is the ratio of the standard deviation
to the mean.
calculates the
corrected sum of squares of a measure. The corrected sum of squares is the sum of the squared deviations from
the mean.
T-statistic (for Average = 0)
calculates the Student’s t statistic
for a measure, assuming a mean value of zero.
P-value (for T-statistic)
calculates the probability
of observing the t statistic
value or a more extreme value. A small value indicates that the mean
is likely not equal to zero.