Calculating Metrics
Metrics are measurements that can go up and down over time. Each metric should map to a measurement of a process or workflow in your organization. Counts of events, revenues and cycle times are good examples of metrics you may want to add to your data model.
You will setup and update metrics from the Add Data Model workflow.
Types of Metrics
Name | Input Column Type | Description |
---|---|---|
row count | No input column required | Count of rows |
custom metric | No input column required | A SQL-based custom aggregation, see Custom SQL Metrics for more details |
count | String or Numeric | SQL count() over a field - counts the number of times the field is not NULL |
count distinct | String or Numeric | SQL count(distinct ) - counts the number of unique, non-null values in a column |
min | Numeric | SQL min() - finds the minimum value of a column |
max | Numeric | SQL max() - finds the maximum value of a column |
avg | Numeric | SQL avg() - finds the arithmetic mean of a column |
The row count metric
The row count metric is added by default to a data model and can't be removed. This is intentional - see the best practices for defining this metric in the Data Modeling Best Practices
Column Aggregations
The simplest type of metric. See table above for which functions you can run on specific metrics.
To add a column aggregating metric:
- Go to the Add Data Model workflow by creating a new data model or editing an existing model
- In the list of fields, click
Add metric
next to the name of the field you'd like to aggregate - Click on the new item in the Metric list to expand the and select the aggregation type.
Custom SQL Metrics
Custom SQL metrics allow you to have more involved custom aggregations in Glean like weighted averages or proportions and other formulas. You can define a custom metric as any aggregating function that results in a numeric value. To add a custom metric:
- Go to the Add Data Model workflow by creating a new data model or editing an existing model
- Under the Metrics section click the
+ Add Custom Metric
button - Enter your aggregation and click the 🔄 button to test the SQL