These are the cross-source questions azimuth answers — the ones a single feed,
or a static OKF bundle, cannot. Each answer connects two or more channels
(energy, geophysics, climate), links every factual claim to its L1 source note,
and names the use-case it serves. Generated live from the 2026-W35 bundle and
regenerated every weekly cycle — the living answer is the point.
Q1
Is Europe's energy supply getting safer or more fragile right now?
EU gas storageUS crude inventoriesEnergy prices
Serves: Energy / policy desk — a one-glance supply-health read across the physical balances and the price tape.
Verdict — the data leans well-supplied, not fragile. Storage is filling and spot crude eased, while the multi-week US crude draw is the single signal to watch; azimuth states what the feeds show, not a safety call (natural-gas-storage-eu, crude-oil-inventories, energy-prices)
EU gas storage is building — 3,169 Bcf as of 2026-08-14, +16 Bcf week-on-week, extending the injection run to 8 straight weeks (natural-gas-storage-eu)
US crude inventories drew down -863 (EIA week of 2026-08-14) to 722,241 (crude-oil-inventories)
Spot crude eased this week — WTI $84.05/bbl (+6.5), Brent $92.51/bbl (+5.3) on the reported week-on-week change, so the price tape is not signalling scarcity (energy-prices)
Show your work — what if the data were different?
Input fed to the verdict: EU gas storage weekly change — +16 Bcf (building)-16 Bcf (drawing down)(sign flipped)
Truth-table branch fired:gas storage buildinggas storage drawing down
Verdict — the data leans well-supplied, not fragile. Storage is filling and spot crude eased, while the multi-week US crude draw is the single signal to watch; azimuth states what the feeds show, not a safety call (natural-gas-storage-eu, crude-oil-inventories, energy-prices)
Verdict — the data leans fragile, not well-supplied. Storage is drawing down — the physical buffer Europe leans on is shrinking, not filling; that is the fragile side of the read; azimuth states what the feeds show, not a safety call (natural-gas-storage-eu, crude-oil-inventories, energy-prices)
The verdict you see is recomputed from the input above by synthesis/answers.py at build time — flip the sign and the branch, and the verdict, change. Nothing is canned.
Q2
Did supply or demand move energy prices this week?
US crude inventoriesEnergy prices
Serves: Economist / market analyst — separating a supply story from a demand story without a paywalled terminal.
Verdict — supply drove the tape. A drawdown and a higher price line up: a coherent supply-tightening week (crude-oil-inventories, energy-prices)
Supply side: US crude stocks tightened — -863 week of 2026-08-14. A draw normally argues for FIRMER prices (crude-oil-inventories)
Price side: WTI $84.05/bbl rose (+6.5 reported w/w) — the actual tape, against what the inventory draw implied (energy-prices)
Show your work — what if the data were different?
Input fed to the verdict: WTI weekly price change — +6.5 (rose)-6.5 (fell)(sign flipped)
Verdict — supply drove the tape. A drawdown and a higher price line up: a coherent supply-tightening week (crude-oil-inventories, energy-prices)
Verdict — demand, not supply, set the tape. Inventories tightened yet prices fell: the bearish price move overrode a bullish supply signal, so the week's driver was softer demand expectations, read straight off the two feeds (crude-oil-inventories, energy-prices)
The verdict you see is recomputed from the input above by synthesis/answers.py at build time — flip the sign and the branch, and the verdict, change. Nothing is canned.
Q3
Did any major earthquake this week put energy infrastructure or population centres at risk?
GeophysicalEnergy supply
Serves: Risk & humanitarian desk — a fast, non-alarmist read of whether a seismic week actually touched the energy map.
Largest recorded event: M6.7 31 km NW of Aniso, Peru — one of 48 events at or above M5 USGS logged this week (earthquakes)
No observed reach into energy infrastructure. The week's quakes cluster away from the physical energy-supply core (US crude inventories, EU gas storage) and from the fuel-reporting countries — the data shows seismicity and the energy balances did not intersect this week (earthquakes, crude-oil-inventories)
azimuth reports what USGS RECORDED, never what may happen next — a sourced 'no significant overlap' is the honest, efficient answer when that is what the week's data shows (earthquakes)
Q4
What is the single biggest shift in the world's open data this week — and what does it connect to?
EnergyClimate signals
Serves: Journalist / newsroom editor — the lede plus its connective tissue, ranked by a transparent rule, not vibes.
Biggest move: WTI Crude Oil, +8.4% week-on-week — the largest swing across the quantitative energy series this week (energy-prices)
What it connects to: the move sits inside the inventories-vs-price loop — US crude drew down while spot prices eased, so the headline swing reflects demand-side repricing rippling from the spot tape into the physical balances and on to pump prices (crude-oil-inventories, energy-prices, fuel-prices)
The slow-moving record: atmospheric CO2 stands at 427.11 ppm (Mauna Loa, 2.83 ppm/yr) — not a weekly 'shift' but the baseline every energy story is told against; the demonstrator flags it as a different time-scale, not the week's headline (co2-monitoring)
Q5
Show me everything that connects a given region or commodity across the data.
Energy supplyGeophysicalClimate signals
Serves: Researcher / analyst — the cross-channel graph for one subject, every edge clickable to its L1 note.
Commodity spine — crude oil ties three feeds together: the physical balance (-863 EIA stocks, 2026-08-14); the spot price (WTI $84.05/bbl); the pump (downstream fuel-price panel). One commodity, traced from the ground to the pump across separate L1 feeds (crude-oil-inventories, energy-prices, fuel-prices)
Papua New Guinea surfaces under 3 channels this week (conflict-watch + Geophysical + humanitarian) — a co-occurrence in the open data, reported as a link, not a cause (conflict-events-ucdp, displacement-flows, earthquakes)
Generated deterministically by synthesis/answers.py from the live
L1 bundle (the latest dated note per editorially-clean source key; held themes excluded).
Numbers are read straight from the source notes — nothing is invented. An honest,
sourced “no significant overlap” is itself a valid answer when that is what the
week’s data shows. Re-run after any ingest and the answers refresh in place.