Description |
Divides one or more time-series by another time-series |
|
Input |
Multiple time-series |
|
Parameters |
None |
|
Output |
Grouped Time Series: Multiple |
|
Available in |
Alerts, Composites, Dashboards |
The ratioPairs function is one of the most commonly used in Anodot. It provides a simple way to divide one (or more) time-series by another time-series. This is often used to generate composites such as CPC (Cost Per Click) and CTR (Click Through Ratio).
Usage example (CPC)
The ratioPairs function is created by entering two metric expressions; one as the dividend, and another as the divisor. The following example shows how to set up a CPC composite, based on the division of cost by click samples, that are grouped and matched by their country property:
A more detailed example
In Anodot, we use the ratioPairs function to calculate the usage of metrics in each of our customers’ accounts. This is done by dividing the number of indexed metrics (metrics) by the total number of contracted metrics (max_metrics):
The expression below returns the result of dividing each dividend time-series sample by the matching sample in the divisor time-series. The GroupBy function is mandatory here, as the two groups of metrics must be matched in at least one dimension. In our example, this allows the one-to-one matching of metrics and max_metrics samples that share the same customer and stack properties. The result of this composite is multiple time series, each one expressing the ratio between metrics and max_metrics for a specific customer and stack combination.
For another detailed example, check the following tutorial.
Note: the Pairs function is similar to ratioPairs, but offers additional operations between matching time-series samples (multiplication, addition, and subtraction).