<instance_name>!GET_DRIVERS

Finds the most important dimensions in a dataset, builds segments from those dimensions, and then determines which of those segments most influenced the metric.

GET_DRIVERS is well-suited to extracting root causes from datasets that have a large number of dimensions. Continuous dimensions are also supported without pre-processing them into categorical dimensions, and the results can indicate dimensions with negative conditions (for example, “region is not North America”).

If you need to select specific columns from the data returned by this method, you can call the method in the FROM clause of a SELECT statement. See Selecting columns from SQL class instance methods that return tabular data.

Syntax

<model_name>!GET_DRIVERS(
  INPUT_DATA => <input_data>,
  LABEL_COLNAME => '<label_colname>',
  METRIC_COLNAME => '<metric_colname>'
);
Copy
INPUT_DATA

A reference to a table, view, or query. All columns other than the ones specified by LABEL_COLNAME and METRIC_COLNAME are taken as dimensions to be considered by Top Insights. Numeric columns are taken to be continuous dimensions, while string and Boolean columns are considered categorical dimensions. To treat a numeric column as a categorical dimension, cast it to a string.

LABEL_COLNAME

The name of a Boolean column in INPUT_DATA designated as the label that indicates control data (FALSE) vs test data (TRUE).

METRIC_COLNAME

The name of a FLOAT column in INPUT_DATA representing the value of interest that has been influenced by the included dimensions.

Output

Column

Type

Description

CONTRIBUTOR

ARRAY

ARRAY of strings describing a segment or insight from the algorithm.

METRIC_CONTROL

FLOAT

The total value of the metric in the control period in a specific segment.

METRIC_TEST

FLOAT

The total value of the metric in the test period in a specific segment.

CONTRIBUTION

FLOAT

The absolute impact of a driver on the overall metric change. The higher the value is, the higher the associated absolute impact on the overall metric change.

RELATIVE_CONTRIBUTION

FLOAT

The percentage impact of a metric change. The higher this value is, the higher the associated relative impact on the overall metric change.

GROWTH_RATE

FLOAT

The percentage of a metric change.

Usage Notes

  • Execution time scales with the number of dimensions and the cardinality of those dimensions.

  • The input metric must be an individual observation or an aggregate.

  • For categorical dimensions having more than 25 values, Top Insights uses only the top 25 most influential values to create segments.

Examples

See Examples.

Language: English