Filters

Global filters let you narrow all widgets on a dashboard to a specific subset of your data — for example, only events from a particular platform, app version, or country — without touching any individual widget query.

How it works

When one or more global filters are active, the backend wraps every widget’s table reference in a pre-filtered subquery:

-- Without filters
SELECT ... FROM <events_view> WHERE ...

-- With a global filter (e.g. platform = 'steam')
SELECT ... FROM (SELECT * FROM <events_view> WHERE platform = 'steam') WHERE ...

This happens transparently — your widget SQL stays unchanged. The filter is applied before your SELECT and WHERE clauses run.

Global filters stack with the dashboard time range: both are applied simultaneously at the data layer.

Filter properties

You can filter on any top-level column from the events schema:

Column	Example value
`name`	`app_started`
`platform`	`steam`
`app_version`	`1.7.9`
`country`	`France`
`region`	`Nouvelle-Aquitaine`
`identity`	`<user UUID>`

You can also filter on nested fields inside the value or custom JSON columns using dot notation:

Property	Example
`custom.plan_type`	`premium`
`value.screen`	`main_menu`

Operators

Operator	Behaviour	Example
`=`	Exact match	`platform = 'steam'`
`!=`	Not equal	`platform != 'ios'`
`>`	Greater than	`app_version > '1.7.0'`
`<`	Less than	—
`>=`	Greater than or equal	—
`<=`	Less than or equal	—
`LIKE`	Contains (auto-wrapped with `%`)	`country LIKE 'United'` → matches `United States`, `United Kingdom`
`NOT LIKE`	Does not contain	—
`IN`	Matches any in a comma-separated list	`platform IN 'steam, ios'`
`NOT IN`	Matches none of the list	—

Multiple filters

When you add more than one filter, they are combined with AND — all conditions must be true simultaneously.

Example: Only events from Steam on version 1.7.9:

Property	Operator	Value
`platform`	`=`	`steam`
`app_version`	`=`	`1.7.9`

This is equivalent to adding WHERE platform = 'steam' AND app_version = '1.7.9' before every widget query on the dashboard.