A research system that treats FactIQ as one data source among several, never a single point of failure - built around one question: is the data point-in-time aware? It is confirmed not vintage-aware (see the vintage probe), which shapes everything else in this project.
Testing whether official Asian trade/industrial statistics diverge from independent, satellite and shipping-derived reality - and where they do, whether that divergence is tradeable, or just noise. Built for early divergence signal detection between official statistics and observed physical/market reality (satellite, shipping, trade-mirror data), with strict point-in-time discipline throughout.
Official industrial output vs. a satellite composite; Yunnan dry-season hydropower curtailment signal. H1 weak, H2 not falsified.
Read the finding →Germany anonymizes all gas-partner data; Belgium↔Netherlands mirror statistics fail to reconcile.
Read the finding →Semiconductor export co-movement - the pipeline reproduces a well-documented relationship (r=0.63).
Read the finding →Capex language never net-positive while supply language always is, every quarter, both majors checked.
Read the finding →"Waste" oil priced at parity with virgin palm oil in Spain - not the discount genuine waste should carry.
Read the finding →Statistically unusual points, ranked and cited, awaiting human review.
Read the finding →Password required.