How a number gets onto this site. The pipeline view is the system: APRA releases and fund disclosures flow through registries, immutable archives and QA gates into the warehouse, then through the semantic layer — where citations, masking and comparison rules are enforced in code — to the serving database. The metric-lineage view is generated live from the served metric registry: every measure the site charts, traced from its source table to the sections that use it.
sourceprocessQA gatestoresemantic layermodelled (estimates)in warehouse, not yet surfaced
APRA landing pages
6 publication families · XLSX / CSV / ZIP releases
33 curated of ~8,900 ids · drafted serve values only
Comparison rulesLOCKED
refusals live in code, not prompts
Query serviceLOCKED
every value cited · masked stays masked
Eval
64 cases · 100% required to ship
Modelled artifactestimates
allocation × ETF proxies · MAE published · never in reported tables
Neon Postgres
9 read-only serving tables, rebuilt per export
MCP server
thin skin over the service · 6 tools
Super Fund Analytics
12 sections · table-level source lines on every chart
Press enter or space to select a node. You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel.
Press enter or space to select an edge. You can then press delete to remove it or escape to cancel.
Pipeline topology is authored alongside the code it describes; metric lineage is generated from the serving database's metric table and the section constants — an edge exists only where a section actually charts that measure.