Analytics
Drop a CSV, get a typed governed managed table — classified, masked, queryable from workflows, notebooks, and the ontology layer.
Scrydon Analytics is the catalog + warehouse + governance layer that lets any team ingest, classify, and query structured data without standing up a separate data platform. Upload a file, get back a typed governed table — with row-level access, column masking, classifications, profiles, and one click to a notebook.
The mental model is Foundry-style data catalog on top of an OLAP warehouse: drop a CSV, the platform infers types, applies classifications, and exposes the table through the same authorisation and audit layer the rest of Scrydon uses.
Table lifecycle
Classification and masking are enforced at every read — workflows, notebooks, and raw SQL all share the same governance path.
What you can do
| Capability | Read more |
|---|---|
| Drop a CSV / JSON / JSONL and get a typed table | Managed tables |
| Lossless preservation of dotted / non-SQL-safe headers | Column names |
| Classify columns and apply masking policies | Classification & masking |
| Query from workflows, notebooks, or raw SQL | Querying |
| Let an agent create its own tables | Agent-created tables |
| Run a Python notebook against your tables | Marimo notebooks |
How it relates to the rest of Scrydon
- Workflows read managed tables through the
scrydon:tablestools — get-schema, query, write, delete. - The ontology layer projects typed Objects on top of managed-table rows. Same data, but typed and traversable. See Ontology.
- Security is enforced at every read: column masking, row filters, and audit logging come for free, regardless of caller.
Where Analytics lives in the cluster
Analytics is one of the three top-level product surfaces (Platform / Agentic / Analytics), running in the scrydon-analytics namespace. The architecture is documented at Architecture → Analytics stack.