Research · Methodology

How Journey sources, connects and labels data

Every number Journey AI by Journey Foods shows should have a source, a date and a clear label when it is an estimate. This page explains how.

Short answer: Journey AI by Journey Foods is connected data infrastructure. It links product, ingredient, supplier, packaging and price data into one graph, and our standard is that every record carries its source and date. Data comes from public datasets, licensed datasets and integrations, supplier documents, and what each customer loads into their own workspace. We convert units, match names to one ingredient identity, score ingredients on nutrition, cost and sustainability, and show gaps instead of filling them. On public pages, every number is labeled live, model, illustrative or vendor-stated, so people and AI agents can tell what kind of number they are reading.

Price data

The Journey Foods Price Terminal publishes daily exchange reference prices for commodities such as cocoa, sugar, coffee, wheat, corn, soybeans, soybean oil, rice and vanilla, with 30-day model forecasts, by REST and MCP. The JAI food price outlook models public price series and is tested with rolling-origin backtests against “no change” and trend baselines before release. We will publish those results with the method once they are approved.

The data objects

  • Products

    Finished goods with their ingredient statements, packaging and where they are sold.

  • Ingredients

    Function, nutrition, allergens, certifications, sustainability context and the products that use them.

  • Suppliers

    Who makes each ingredient, origin, and the documents a supplier has provided.

  • Packaging

    Materials and formats linked to products.

  • Prices

    Dated commodity reference prices and model forecasts from the Journey Foods Price Terminal.

  • Relationships

    The links between all of the above. The graph is the product, not any single table.

From source to answer

  1. 01

    Ingest with source and date

    Our standard: every record carries where it came from and when.

  2. 02

    Normalize and connect

    Units are converted (for example to USD/kg), names are matched to one ingredient identity, and relationships are built.

  3. 03

    Score and expose

    Journey AI scores ingredients on nutrition, cost and sustainability, shows gaps, and serves the same data in the app, by API and through connectors for AI agents.

Labels we use on public pages

  • Live: read from a Journey feed at render time; the date is shown.
  • Model: a forecast or score produced by a Journey model; never presented as an observed value.
  • Illustrative: a teaching example; not a customer result and not market data.
  • Vendor-stated: a figure taken from another company's own public page, linked, with the date we read it.

What we do not publish

Customer results without approval, testimonials we cannot attribute, time-saved or ROI percentages without a measured study, and claims about competitors that their own public pages do not support.

Questions buyers ask

Data provenance

This page describes Journey's own practice. It will be updated when benchmark results are approved for publication under /research.

Methodology: how Journey sources, scores and labels data.

Written by Riana Lynn and the Journey Foods team. Last reviewed . Vendor details change; tell us at hello@journeyfoods.io if something here is out of date.

See it on your own formula.

Book a demo and the Journey Foods team will run one of your formulas through Journey AI, or start a trial and load it yourself.

Book a demo