Demand forecasting & aircraft intelligence
Forecast. Optimize. Decide.

Know the demand before you fly the route.

Route-level passenger and cargo demand forecasts to June 2027 across 125,000+ directional airport pairs — built entirely from official national aviation statistics, with per-route confidence classes and the model reasoning shown, not hidden.

Directional routes
125,667
Historical route-months
2.28M
Forecast horizon
JUN 2027
Regions covered
09
Forecast · Optimize · Decide

Three layers of decision support, one workflow.

From raw demand history to a recommended aircraft and frequency — each step is transparent and exportable.

01 · Forecast

Demand forecasting

Passenger and cargo demand, historical and forecast, at route level through June 2027 — every forecast carries a 95% confidence interval and a per-route confidence class, so you see the uncertainty, not just a point estimate.

PAX · CARGO · ROUTE-LEVEL
02 · Optimize

Route & market intelligence

Traffic trends, load factors, carrier activity, and market concentration across major corridors — with forecasting diagnostics and anomaly monitoring you can trust.

LOAD FACTOR · HHI · CARRIER MIX
03 · Decide

Aircraft & frequency

Demand-driven aircraft suitability and frequency guidance from a reinforcement-learning agent, paired with transparent operational logic — auditable, not a black box.

SUITABILITY · FREQUENCY · BELLY-CARGO
Who uses it

Built for the people who plan the network.

Airline planning teams

Network, fleet, and cargo planners sizing routes and testing frequency and aircraft assumptions against real demand.

Consultancies & analysts

Advisory and research teams needing defensible, source-transparent forecasts they can put in front of a client.

New entrants & investors

Teams evaluating routes, aircraft, and market opportunities before committing capital or capacity.

Data coverage

Additive coverage, built source by source.

Each region is ingested from official government sources — no scraped or proprietary black-box feeds. 2.28 million historical route-months across 125,667 directional airport pairs. Where a country does not publish route-level statistics, its routes appear only as reported from the other end, and we say so rather than presenting partial coverage as complete.

Domestic and international routes

North America — US, Canada Europe — EU 27, UK LATAM North — Mexico LATAM South — Brazil, Argentina, Colombia, Chile, Peru Oceania — Australia, New Zealand

International routes only — reported from the counterpart authority

Asia Middle East Africa Antarctica

Sources: US BTS T-100 · Eurostat · UK CAA · Statistics Canada · AFAC Mexico · ANAC Brazil · ANAC Argentina · Aerocivil Colombia · JAC Chile · MTC & DGAC Peru · BITRE Australia · Stats NZ.

Questions

What planners ask before they trust a forecast.

Where does the traffic data come from?

Every historical figure comes from an official national statistics authority — BTS T-100, Eurostat, UK CAA, Statistics Canada, AFAC, ANAC, Aerocivil, JAC, MTC, BITRE, Stats NZ. Nothing is scraped or bought from a black-box feed. Where two authorities report the same route, both records are kept and precedence is resolved at read time, so the source of every number stays traceable.

Which regions and countries are covered?

Domestic and international routes for North America, Europe, LATAM North, LATAM South and Oceania. Asia, the Middle East and Africa are covered for international routes only — those reported from the other side of the route — because the countries concerned do not publish route-level statistics.

How far ahead does the forecast run?

To June 2027, covering horizons of one to twelve months. Horizons beyond twelve months are deliberately not published: in backtesting they lost to a seasonal naive benchmark, so publishing them would add noise rather than information.

How accurate are the forecasts?

Accuracy is measured by rolling-origin backtesting at three separate forecast origins, so no single lucky cut-off can flatter the result. Every route carries a published confidence class — High, Moderate, Low or Unrated — derived from its own measured error, and every forecast shows a conformal 95% confidence interval calibrated on held-out data. Thin and unstable routes are labelled, not quietly presented as reliable.

Is air cargo covered as well as passengers?

Yes — forecast separately across roughly 16,000 cargo routes, combining freight and mail tonnage. Belly capacity carried on passenger aircraft is netted out first, so the signal reflects residual demand needing dedicated freighter capacity rather than double-counting what already flies in the hold.

What aircraft and frequency guidance is given?

Aircraft suitability is scored by a transparent heuristic across range, gauge, fuel burn and airframe age, with each candidate labelled by generation so you see the trade-off instead of having older types silently filtered out. Frequency guidance comes from a reinforcement-learning agent trained on a newsvendor cost that weights spilled demand above empty seats, and scored against outturn traffic rather than the forecast used to set capacity.

Can the data be exported?

Yes. Historical series, forecasts with confidence intervals, and carrier-level breakdowns export to CSV on any plan — no seat limits, no per-export charges.

Pricing

One plan. Full access. No seat games.

Everything is included from day one — forecasts, recommendations, exports, and every live region. Choose how you pay.

Monthly
$49
per month
  • Full platform access
  • Demand forecasts + 95% CI
  • Aircraft & frequency recommendations
  • Full data exports
  • Model intelligence dashboard
  • All current regions
Get started
AnnualSave ~17%
$490
per year
  • Everything in monthly
  • Two months effectively free
  • Priority on new-region access
  • Full data exports
  • Model intelligence dashboard
  • All current regions
Get started →

See the demand before you commit the capacity.

Airline planners, consultancies, and analysts use the Aviation Intelligence Platform to forecast demand and pressure-test aircraft and frequency decisions.

Launch the platform →