About the trend type of the Trend component

The trend type of the Trend component detects either consistently high upward or downward changes in data values over a length of time. The length of time is fully divided into equal periods for which some calculated value in each period is either consistently increasing (or consistently decreasing) by a certain percentage, period to period across all periods.

To detect a trend, you must specify a field in the data source on which to do the trending. You must also specify the time period over which the trending is done and the percentage increase or decrease you are looking for. The data values all come from a designated field in a pre-populated container or select function.

The trend groups data values by fixed time intervals within the specified period. (For example, if a period spans three calender months beginning January 1 to the end of March, the data values are grouped into intervals for the months of January, February, and March.)

If the periods do not have subgroups (each period is not further divided into subgroups), then they are evaluated with a mathematical function (such as sum or average for example), and the resulting values are compared and the trend returns true if the percentage change of the values is consistently above the threshold value.

If you choose to use subgroupings within the trend period, Detect divides the trend period into subgroup time blocks, applies the subgroup function to the data in each subgroup, returns a value for each, and only then applies the selected trend period function on that set of returned values.

In order for the component to detect an upward trend, set the trend percentage value to a positive value. To detect a downward trend, set it to a negative value.

Example of a trend

The options for the trend