Define your metrics once. Use them everywhere.
Golden Analytics reads Apache Ossie semantic models, so the business definitions your team agreed on travel with your data into every analysis, dashboard and AI answer.
An open standard for business meaning, plugged into AI-native BI
Bring the model you already have
Apache Ossie (Incubating) is an open, vendor-neutral format for business semantics: datasets, fields, metrics, relationships and the context around them. Upload your Ossie YAML to a dataset in Golden. No proprietary semantic layer to rebuild, no lock-in.
Checked against your real data
Golden validates the model before it lands. For a warehouse dataset it checks the live warehouse schema; for an uploaded file it checks the file's columns. You get a report of which tables and fields matched, which metrics translated, and why any metric was skipped. There are no black boxes.
One definition, every surface
Matched fields and metrics are applied to the dataset, so charts, dashboards and stories use the same definitions your data team approved. Edit the model as YAML, re-validate, and the change is re-applied.
What Golden does with an Ossie model
| Step | What happens |
|---|---|
| 1 · Upload | Open a dataset's Semantic model tab and upload an Ossie YAML file, or add one while importing tables from a warehouse. Golden reads Ossie spec version 0.1. |
| 2 · Parse | Golden reads datasets, fields (dimensions and facts), metrics, relationships, synonyms and AI context. Files up to 2 MB, with up to 500 datasets and 5,000 each of fields, metrics and relationships. |
| 3 · Validate | Each table and field is matched against the dataset's real columns. A model that describes none of the dataset's tables is refused instead of half-applied. |
| 4 · Translate | Metric expressions are translated into Golden formulas. Anything Golden can't safely translate yet, like a cross-table metric, is listed with the reason rather than silently dropped. |
| 5 · Apply | Fields and metrics enrich the dataset. One model per dataset, stored with access controls, and only dataset editors can change it. |
Why Golden with Apache Ossie is different
| Criterion | Traditional BI | Golden Analytics |
|---|---|---|
| Metric definitions | Rebuilt inside each tool's own semantic layer | Read from the Ossie model your team already maintains |
| Portability | Definitions are locked to one vendor's format | An open, vendor-neutral YAML format you can take anywhere |
| Validation | Mismatches show up later as wrong numbers | Every table, field and metric is checked against your data before it applies |
| Untranslated metrics | Dropped or approximated without saying so | Listed in the validation report with the reason |
| AI answers | AI guesses what a metric means from column names | AI uses the definitions, synonyms and context in your model |
Golden connects to the rest of your stack too
Snowflake
AI analytics on the Snowflake data, roles, and metrics you already run the business on. Reads your Semantic Views; OAuth means Snowflake roles decide access.
Databricks
AI analytics on the lakehouse data, Unity Catalog grants, and metrics you already run the business on. Every statement lands in your Query History.
Alteryx
Alteryx prepares the data. Golden makes it sing. Curated data that took hours to prepare, in front of every analyst and business user in minutes.
Golden × Apache Ossie
What is Apache Ossie?
Apache Ossie (Incubating) is an open specification for describing business semantics, such as metrics, dimensions and the relationships between tables, in a portable YAML format. It came out of the Open Semantic Interchange effort and is now incubating at the Apache Software Foundation. Golden Analytics is part of the Ossie ecosystem.
Which parts of an Ossie model does Golden use?
Golden reads datasets, fields, metrics, relationships, synonyms and AI context from Ossie spec version 0.1. Fields and metrics that match your data are applied to the dataset. The validation report shows exactly what matched and what didn't.
What happens to metrics Golden can't translate?
They're listed in the validation report with a plain reason, for example a metric that spans several tables, or one that uses columns the dataset doesn't have. Nothing is dropped silently, and the rest of the model still applies.
Does Golden send my data anywhere to read the model?
No. Validation compares the model against your schema: table and column names. For warehouse datasets Golden queries the warehouse in place, and no customer data is sent to AI models.
Do I need Ossie if my metrics already live in Snowflake or Databricks?
No. Golden consumes Snowflake semantic views and Databricks metric views directly, so those definitions arrive in Golden as they are. Ossie is how you bring semantics that live anywhere else, in a portable, open format.
Bring your Ossie model to Golden
Upload your semantic model and have governed metrics in a working dashboard in minutes.