INDUSTRY FIELD GUIDE
Coal and mineral wholesale sampling and analysis basis for AI
Two reports can show similar values while expressing results on different moisture or analysis bases.
NAICS 423520 · Coal and Other Mineral and Ore Merchant Wholesalers
Two reports can show similar values while expressing results on different moisture or analysis bases. USGS publishes coal quality data that includes chemical and physical measurements. A wholesale exercise can test whether a model preserves the stated basis and knows when to request a missing input. USGS coal quality data
Invent two sealed dummy shipment lots, a sample receipt and a pair of assay summaries. Give one an explicit basis and one an unspecified basis. Ask which comparison is supported and how the report traces to a particular lot. Score reference and unit independently.
Use inert props and abstract values. Exclude actual commodity samples, confidential mine information, supplier contracts and regulated handling details.
Describe a fictional mineral assay records sample
Original Fiund exercise
An assay result needs its sample, lot and analysis basis before comparison
Two sealed inert props and original assay summaries use fictional analyte and lot aliases. No actual mineral, mine, proprietary method, regulated material or handling instruction appears.
| Record | Source | What it records |
|---|---|---|
| MAW1 | Lot A chain | Sample S-A received from lot A. Abstract analyte X reports 12 practice percent on a clearly labeled dry basis. |
| MAW2 | Lot B chain | Sample S-B receipt explicitly matches lot B. X also reports 12 percent, but analysis-basis field blank and no conversion input. |
| MAW3 | Loose report | Same lot-A-style title but sample S-C and different unit U. No approved cross-reference to A or B. |
Question for the model
Which sample-to-lot links are supported, and do the equal 12-percent values establish comparable assay results?
Show the reference answer and checks
MAW1 and MAW2 independently link S-A to A and S-B to B, so identity is supported for both. Only A states dry basis. Equal 12-percent values do not establish a like-for-like comparison without B's analysis basis and any needed conversion inputs. MAW3 is unrelated by sample and unit. Preserve proven lineage and separate comparability gaps rather than infer moisture basis or merge results because their values look alike.
- Comparing equal percentages without matching the underlying analysis basis.
- Substituting an unlinked report from a similarly titled lot.
Fields to preserve
- Lot and sample receipt
- Supports lineage.
- Analyte, unit and analysis basis
- Defines the measure.
- Conversion input and missing scope
- Limits comparison.
Test another case
Supply B's dry-basis confirmation but switch its analyte code to Y. Require basis agreement without treating different analytes as comparable.
Use original inert props and abstract values. Exclude real assays, mine information and supplier agreements.
Discuss a contribution
Describe the sources you control, the task they could support and what has already been checked. Use the existing-data assessment for an archive or the collection brief for a new recording project. The contributor agreement defines Fiund’s licensing authority and compensation. A guide or fictional exercise does not establish an accepted collection or a sale.