●

About the India Trade Terminal

What this is, how it is put together at a conceptual level, and where the line sits between government data and our own work. For the calculation and signal formulas, see /methodology.

What this is

The India Trade Terminal is a research tool for India's official merchandise trade statistics: every HS code from chapter down to 8-digit tariff line, every trading partner country, monthly since April 2018. It exists to make that public data genuinely usable — searchable, comparable, screenable, and checked for unusual changes — rather than left as raw government tables.

Data sources

Three government releases, published at different speeds and levels of detail. They are kept as separate layers and never blended; every page labels which one you are looking at and for which month.

1. Detailed trade data · DGFT TradeStat

The permanent detailed dataset: country × HS code (chapter to 8-digit tariff line), monthly since April 2018.

  • Permanent historical analytics, screener results, signals, anomaly detection and backtests use this layer.
  • The newest month is provisional in DGFT's own release and can be revised.
2. Selected-Country Quick Estimate · DGCI&S

The latest provisional country-level month, published ahead of the detailed data, for selected partner countries only.

  • Supports only limited, clearly-labelled provisional calculations for the newest country-level month (value, YoY, MoM, FYTD, share, covered-member group totals and member drivers).
  • No HS/commodity breakdown, and it does not feed the permanent detailed analytics, signal engine or backtests. A country missing from the release is not covered -- not zero trade.
  • Superseded automatically once DGFT's detailed data for that month lands.
3. India Topline Estimate · Department of Commerce (monthly provisional press release)

India's total merchandise exports and imports for the latest month -- India-level only.

  • A separate provisional all-India layer: no country or HS breakdown, and separate from the Selected-Country Quick Estimate.

Details, lags and the exact labels: /methodology#data-layers.

How it is built

At a conceptual level, one pipeline produces the permanent historical analytics, screener results, signals, anomaly detection and backtests, all from the detailed DGFT TradeStat layer:

Detailed source data (DGFT TradeStat) | v Validation | v Analytical marts | v Deterministic analytics | v Signals | v Research interface (this terminal)

DGCI&S Selected-Country Quick Estimates support only limited, clearly-labelled provisional calculations for the newest country-level month. They do not provide HS-level detail and do not feed the permanent detailed analytics, signal engine or backtests. The India Topline Estimate remains a separate provisional all-India layer.

Each stage only reads the validated output of the one before it. Analytics (growth rates, shares, concentration, rank, CAGR) and signal detection (spikes, collapses, share shifts, concentration changes, balance reversals) are both fully deterministic — the same input data always produces the same numbers, computed by ordinary code, never estimated or generated by a language model. Where an AI-assisted research feature exists on the site, it may only rephrase numbers that were already computed this way; it is never allowed to calculate a figure or invent one that these layers didn't already produce.

This description is intentionally conceptual. We don't publish server names, database file layout, deployment scripts, or infrastructure detail here — that isn't useful to a researcher reading this page, and this site is a public target for automated scraping (see /terms).

Source attribution

Theirs: the underlying data. The raw trade figures — what was exported or imported, of what commodity, with which partner country, in which month — originate from Government of India publications: DGFT TradeStat (Directorate General of Foreign Trade) for the detailed data, DGCI&S for the Selected-Country Quick Estimate, and the Department of Commerce for the India Topline Estimate (see Data sources above). We do not claim ownership of that underlying government-published data, and we do not alter its substance.

Ours: everything built on top of it. That includes cleaning and structuring the raw monthly releases into a consistent historical record; validating and building the analytical marts the permanent analytics read from; the deterministic analytics engine (growth, shares, concentration, CAGR, and the rest — see /methodology#calculations); the signal detection engine and its published historical backtest (see /methodology#signals and /signals/validation); and the research tooling and presentation you are using right now. This is consistent with how /terms frames it: the source data is government-published and not ours; the differentiated, derived layers built on top of it are the specific thing this product is and the specific thing its usage terms protect.

Interactive access vs. bulk data

Using the terminal interactively — looking up commodities and countries, running the screener and comparisons, reading signals, exporting the results of a page you're looking at — is public, free, and exactly what this site is built for.

Automated bulk extraction, systematic scraping, or reconstruction of the underlying dataset from site output is a different matter and is restricted. It is not documented here as an API reference or endpoint list — this page is a statement of position, not a specification. The actual policy, including what counts as normal use and what requires prior permission, is published at /terms.

Research disclaimer

Historical trade analytics on this site are for research and informational purposes. Signals describe unusual changes already present in reported trade data — they are not investment recommendations, forecasts, or trading advice. See /methodology#limitations for what the site's own historical backtest actually shows about how reliably a signal tends to persist, and /terms for the full notice.