Categories of Data Sources
Which categories of source feed the analysis?
The analysis draws on six categories of data sources, with the vast majority of inputs coming from public official sources.
Private sources — including private trading platforms, producers' data, and licensors' publications — are used only with permission from the rights holders.
Internally developed mathematical models are not a data source: they fill gaps when a source lags the reporting calendar, as described in the source timing and coverage article.
What comes from government agencies and statistics bureaus?
Government-reported trade statistics form the foundation of the analysis. These official records are collected on a monthly cycle and processed to generate updated assessments. Because they are published by national authorities and statistical bureaus, the figures are public, verifiable, and free of any single party's commercial interest. Processing applies impartial, auditable methods so that each result can be traced back to its official source.
When are multilateral organizations and market exchanges used?
Data from international and multilateral organizations supplements government trade statistics where broader context or technical reference points are needed. Technology licensors form their own source category, contributing published consumption figures, investment levels, and plant capacities that are verified before use. When official data lags behind the reporting calendar, internally developed mathematical models create preliminary estimates. Those estimates are clearly provisional and are replaced with final figures once the official data becomes available.
Does Intratec run its own surveys?
Only in very exceptional circumstances. Intratec is not a primary source for pricing data: the vast majority of inputs come from public official sources, with private sources and producers' data used only with permission from the rights holders. This approach minimizes subjective influence and self-reporting bias in the assessments. Intratec also maintains no commercial relationships with commodity market participants, and all data is processed through transparent and reproducible methods.
When multiple sources cover the same data, which wins?
Official and recognized institutions take priority for final figures; accuracy is favored over speed, and timing gaps are bridged by clearly labeled model estimates until official data arrives.
When more than one source covers the same data point, official and recognized institutions — national statistics bureaus, governmental agencies, and international multilateral organizations — take priority for final figures. Private sources and producers' data, used only with permission from the rights holders, supplement rather than override official records.
In this prioritization, accuracy is favored over speed: a slower official source outranks a faster private one. Official statistics are typically released one to three months after the period they describe, and the analysis is structured around that reporting rhythm rather than trading reliability for earlier numbers.
The timing gap this creates is covered by internally developed mathematical models. Until the official figures arrive, model estimates stand in and are clearly labeled Preliminary or Forecast, so a reader can always distinguish an interim estimate from a settled official value. Once the official data is released and verified, the estimate is replaced and the figure becomes Final. The meaning of these labels is explained in detail on the Data Status Labels page.
Source reliability is not a one-time judgment — it is assessed continuously. Over time this means:
- More sources: the number of sources feeding the analysis keeps increasing.
- Official-first selection: official and recognized institutions are consistently prioritized.
- Improved validation: validation procedures are progressively strengthened.
- Quick replacement: discontinued sources are replaced quickly so coverage stays continuous.
The logic above is the disclosed principle. The internal parameters used to compare and select among specific sources are not published.
How many sources feed the analysis?
The collection portfolio counts 350+ sources, spanning the six source categories: national statistics bureaus, governmental agencies, international multilateral organizations, commodity exchanges, producers' data, and technology licensors. The portfolio is expanded continuously, with priority given to official and recognized institutions, and discontinued sources are replaced quickly so coverage stays uninterrupted. Pertinent sources are continuously mapped and evaluated for addition. Subscriber feedback also feeds the continuous review of the source base, through the same channels described in the AI and Human Validation article.