What we added: satellite nightlights, port AIS, electricity, and eight new risk gauges

Quantamentry — the data-expansion waves, and what they change.

This is a direct sequel to how we built a 169-country macro platform on free public data. That post laid out the eleven-source foundation. This one is the inventory of what two subsequent data-expansion waves added on top — the alternative and high-frequency series that let the platform nowcast rather than just score annual and monthly fundamentals.

What just shipped

Two ingestion waves, one theme: get closer to the present. The additions, with honest coverage numbers:

Port throughput (IMF PortWatch / Kpler AIS) — daily vessel-transit and port-call tonnage for 133 countries, ~450 days of history. This is the same AIS backbone that powers the shipping-chokepoint board; at the port level it becomes a trade-volume nowcast.
IMF Commodity Terms of Trade (CTOT) — a country-level terms-of-trade index for 165 countries, back to 1990. For a commodity exporter this is one of the cleanest long-horizon reads on external income there is.
Electricity generation (Ember) — monthly generation for 85 countries, the backbone of a real activity pulse.
Uncertainty and geopolitical-risk indices — the Geopolitical Risk index (GPR, 43 countries), Economic Policy Uncertainty (EPU, 20 countries), and the World Uncertainty Index (WUI, 141 countries). Text-derived measures of how much the news flow itself is signalling stress.
Eurostat industrial production and retail — monthly IP and retail-trade volumes for 37 European geographies, including Türkiye, Ukraine, and the CEE bloc — not just the euro core.
Satellite nightlights (World Bank via NASA Black Marble VIIRS) — monthly radiance for ~170 countries, 2019–2025, gas-flaring-masked so oil-field flares don't masquerade as economic light.
Sovereign yields — 10-year benchmark yields expanded from 15 to 35 countries, deepening the sovereign-spread inputs behind the liquidity dimension.
Eight new FRED risk gauges — an EM equity-volatility gauge, 5-year / 10-year / 5y5y-forward inflation breakevens, the Chicago Fed's National Financial Conditions Index (NFCI), the St. Louis Fed Financial Stress Index, plus oil-volatility and gold-volatility series. These populate the global-context board alongside the VIX/DXY/commodity gauges already there.

The unifying move: the original platform scored fundamentals — inflation, rates, debt, governance — on their native monthly and annual cadence. These waves add activity and stress series that update weekly or monthly, close to real time. That's the difference between a scorecard and a nowcast.

It is worth being precise about the two flavours here, because they do different jobs. Activity series — electricity, port tonnage, industrial production, nightlights, terms of trade — measure how much real economic work an economy is doing right now, and they feed the growth side of the picture. Stress series — the uncertainty and geopolitical-risk indices, the financial-conditions and financial-stress gauges, breakevens, oil and gold vol — measure how much fear is priced into the system, and they feed the backdrop that every credibility score is read against. A country can look fundamentally sound on annual data while its activity is rolling over and global stress is rising underneath it; the two families of series are what let us see that divergence as it happens rather than in hindsight.

Why these sources?

Every source above clears the same three-part bar that governs everything we ingest, and it's worth being explicit about it because it is the whole economic model of the platform.

First, free. No per-seat licensing, no five-figure annual contract. Ember, Eurostat, FRED, the World Bank, the IMF public data platform, and the academic uncertainty indices all publish at no cost.

Second, keyless or near-keyless. The fewer credentials in the critical path, the fewer silent-auth-failure modes at 3am. Most of these need no key at all.

Third — and this is the one that actually constrains us — licensed for display and redistribution. It is not enough that data is free to read; we serve it through a product, so we need terms that permit re-use in something we sell. That is exactly why the attribution block at the foot of this post is long: Ember is CC BY 4.0, Eurostat is © European Union / CC BY 4.0, the nightlights are World Bank via NASA VIIRS under CC BY 4.0, and each carries its citation because the licence requires it.

The licensing gate, stated plainly: some sources we ingest — notably raw IMF and BIS series — are not cleared for public redistribution. So we never expose them raw. Those series feed the models internally, but the only things that leave the building are derived indicators — scores, gaps, blended pulses, terms-of-trade changes — computed from them. The public API filters raw single-source IMF/BIS rows out at the query layer. The line we hold is: fundamentals in, derived signals out. It is the difference between redistributing someone's dataset and publishing our own analysis built on it.

That gate is not a nuisance we tolerate; it is the reason the whole product is economically possible. Every commercial macro-data vendor — the ones charging five and six figures a year — carries a redistribution clause precisely to stop a customer from doing what we do. Build your model on their feed and you cannot resell the outputs without a much larger, separate licence. Public data with clean re-use terms removes that ceiling entirely, which is why we are so deliberate about staying inside it: the moment we redistributed a restricted raw series we would inherit exactly the constraint we set out to escape. Keeping the raw/derived line bright is what keeps the platform both legal and cheap to run.

What it changes in the scores

The most consequential addition is the growth pulse. Historically the growth dimension leaned on annual GDP and a leading indicator that the OECD has quietly stopped publishing for most countries. Annual data is a terrible way to monitor an emerging market in real time — by the time a GDP print lands, the turn it describes is two or three quarters old.

The new high-frequency series fix that. Three of them — electricity generation (Ember), port-trade momentum (PortWatch), and industrial production (Eurostat) — now feed a monthly activity pulse that blends into the growth score. The mechanics are: final = 0.7 × base + 0.3 × mean(pulse scores), with the blend renormalising when a pulse is missing for a country. So a country's growth reading now moves when its electricity demand or its port throughput turns, months before the annual accounts confirm it.

We've written the full mechanism up separately — the three pulses, the mapping, and a worked Türkiye example — in nowcasting EM growth with monthly pulses. If you want the how rather than the what, start there.

Beyond growth: the expanded sovereign-yield panel (15 → 35 countries) widens the sovereign-spread input into the liquidity dimension, and the new FRED risk gauges — breakevens, NFCI, the financial-stress index, oil and gold vol — sharpen the global-context backdrop that every country score is read against.

The yield expansion matters more than the headcount suggests. The liquidity dimension leans on the spread between a country's local 10-year yield and the US 10-year as a market-implied read on sovereign stress — the price investors actually charge to hold the paper. Going from 15 to 35 countries with a benchmark yield means that spread signal now exists for a far broader slice of the panel, so more countries get a market-priced fragility input instead of relying on fundamentals alone. And the breakeven series — 5-year, 10-year, and the 5y5y-forward — add a clean read on where inflation expectations sit, which is a useful cross-check on the credibility scores: a bank can be judged credible by the model while the market's own forward breakevens quietly disagree, and now we can see both numbers next to each other.

One honest exclusion up front, because it matters for credibility: satellite nightlights are deliberately not in the growth pulse. The World Bank's monthly VIIRS product publishes with roughly a seven-month lag, which makes it useless for nowcasting. We ingest it and present it as a structural read — a long-horizon activity map, useful for cross-country comparison and for spotting multi-year shifts — but we do not pretend it is timely. Putting a seven-month-lagged series into a "pulse" would be exactly the kind of overclaiming this platform exists not to do.

Where to see it

quantamentry.com/global — the global-context board: volatility and financial-conditions indices, commodities, inflation breakevens, the uncertainty and geopolitical-risk indices, and the shipping-chokepoint stress gauges, all snapshot-dated and refreshed daily.
Country pages — each country's page shows its own inputs, including the activity pulse where coverage exists.
quantamentry.com/methodology — the full data dictionary: every indicator, its source, cadence, coverage, and known caveats.

To watch a handful of countries with the expanded data underneath them, Create a free account — watch 3 countries. Free, and it takes about a minute.

Sources

Port throughput and shipping chokepoints — Sources: Kpler; UN Global Platform; IMF PortWatch (portwatch.imf.org)

Electricity generation — Source: Ember (CC BY 4.0)

Satellite nightlights — World Bank via NASA Black Marble VIIRS, CC BY 4.0

Economic Policy Uncertainty — Baker, Bloom & Davis, "Measuring Economic Policy Uncertainty" (policyuncertainty.com)

Geopolitical Risk index — Caldara & Iacoviello (2022)

Food Price Index — FAO. Food Price Index. Food and Agriculture Organization of the United Nations.

Volatility, financial-conditions, commodity, breakeven, and sovereign-yield series — FRED (Federal Reserve Bank of St. Louis). Terms of trade — IMF Commodity Terms of Trade.

— Quantamentry, July 24, 2026