Identity Time Series Types
Developers might want to produce an instance of a Timeseries Type out of static data. This is done so that static data can be compatible for analysis and machine learning with other time series data that is more time-dependent.
Timeseries Types produced with static data are called identity Time Series Types. Even though this data is constant throughout time, time is forced onto the data to produce the time series.
An example of static data that can use an identity time series is the manufacturer of a SmartBulb object. A SmartBulb manufacturer is a constant characteristic of a SmartBulb object that doesn't change with time.
| time | manufacturer |
|---|---|
| 12:00 | Phillips |
| 12:15 | Phillips |
| 12:30 | Phillips |
| 12:45 | Phillips |
In the table above, the value manufacturer field remains constant over each time interval.
Identity time series data
Producing time series out of static data on the C3 Agentic AI Platform is simple with the help of metrics.
Data is usually modeled as a time series before being passed to a metric to produce an instance of a Timeseries Type. Static data is not modeled as time series before being evaluated by a metric.
When creating an identity Timeseries, you need to specify the source type, and the expression logic to apply on that source type.
For example, the metric below can create an identity Timeseries out of the manufacturer data for SmartBulb objects where the manufacturer name starts with g. The source type (srcType) is smartBulb, and the manufacturer data you want to analyze is found by navigating to the manufacturer field, and then to the name field.
{
"id": "ManufacturerNameStartsWithG_SmartBulb",
"name": "ManufacturerNameStartsWithG",
"description": "True or false for duration of LightBulb",
"srcType": "SmartBulb",
"expression": "startsWith(lowerCase(manufacturer.name), 'g')"
}