Skip to main content

Filters

Maps to: filters

Filters narrow the respondents included in a query — for example, to a demographic segment or an audience defined by their answers. Each query carries a filters list; send an empty list ([]) for no filtering.

How filters combine​

  • Filters in the list are combined with AND — a respondent must satisfy all of them.
  • Multiple values within a single filter are combined with OR — the respondent must match any of that filter's values.

So [age: 18–24 OR 25–34] AND [region: London] means "18–34 year-olds who live in London".

Operators​

Value-based filters accept an op: ==, in, >, >=, <, <=. If omitted, == is assumed.

Filter types​

Look up the available values for each type via the filters catalogue.

Demographic filter​

A standard demographic variable and one or more numeric value codes:

{ "name": "gender", "op": "==", "values": [1] }

Profiles filter​

A Profiles data-series variable. definition_name is <dataseries>__<definition>, e.g. profiles_us__age:

{ "definition_name": "profiles_us__age", "op": "in", "values": ["18-24", "25-34"] }

Metric filter​

Restrict to respondents who gave a particular answer on a metric, optionally for a specific brand_id:

{ "metric": "consider", "op": "==", "values": [1], "brand_id": 1007 }

Composite filter​

A saved, reusable filter referenced by its identifier, together with the specific response options to include:

{ "uuid": "8ab3…", "responses": ["1f0e…", "2a7c…"] }

Raw expression filter (advanced)​

A free-text expression. Powerful but easy to get wrong — prefer the structured types above unless you have a specific need.

{ "expression": "bixdemo_age2 > 1" }

Coming from the BrandIndex UI?​

The UI calls filters "audiences". They map to the API filter types like this:

UI audienceAPI filter type
Core (basic demographics: age, region, gender, income)Demographic filter
Category (variables specific to the chosen category)Demographic filter
Metric (respondents by how they answered a metric)Metric filter
Composite (saved, reusable audiences)Composite filter
Profiles (YouGov Profiles variables)Profiles filter

The UI joins all selected audience variables with AND; to combine variables with OR the UI uses its Filter Builder to produce a composite filter — the same AND-across / OR-within behaviour described above.

Interaction with the scoring population​

Filters and the scoring population both affect who is counted, but they are distinct: filters restrict which respondents enter the query at all, while the scoring population decides who forms the base of the score within that filtered group.