INDUSTRY FIELD GUIDE
Broiler and turkey water-meter data for AI review
A meter reset can appear as a sudden drop in a poultry house’s cumulative water reading.
NAICS 112320 · Broilers and Other Meat Type Chicken Production
NAICS 112330 · Turkey Production
A meter reset can appear as a sudden drop in a poultry house’s cumulative water reading. A reviewer needs to understand the equipment event before treating the graph as a change in the flock. USDA’s Poultry Industry Manual provides production-system context for multiple types of poultry. An invented meter log can test whether an AI assistant reads the chart and event record together. USDA APHIS Poultry Industry Manual
Generate two simulated traces with explicit units. Mark a documented replacement or reset in one and an unexplained reporting gap in the other. Provide house and flock cycle aliases, each with an associated maintenance note. Ask the model what the evidence supports and where a human should look next. Keep age or cycle stage separate when comparing birds and production systems.
Exclude integrator terms, real house plans, actual farm telemetry, and active health information. Use fabricated records and require a qualified operator to review the expected labels.
Describe a poultry event review dataset
Original Fiund exercise
A cumulative-meter reset is not negative water consumption
Two imaginary poultry-house aliases have simulated meter records. No live telemetry, flock-health information, house plan or production recommendation is included.
| Record | Source | What it records |
|---|---|---|
| PW1 | House A, cycle C1 | Cumulative meter at 08:00=500 practice liters, 09:00=5. Maintenance card confirms reset to zero at 08:30, after a verified 510-liter pre-reset reading. |
| PW2 | House B, cycle C2 | Cumulative readings 08:00=200, 09:00 missing, 10:00=218. No reset record or 09:00 observation. |
| PW3 | Cycle metadata | A and B are different cycle stages. No stage-normalization or bird-count basis supplied. |
Question for the model
What interval use can be reconstructed for A, and what limits apply to B's gap and an A-to-B comparison?
Show the reference answer and checks
PW1 supports ten practice liters before the reset, from 500 to 510, and five after it, for 15 across the interval. The raw 500-to-five subtraction would misread the equipment event. PW2 supports an 18-liter cumulative difference over two hours but no observed 09:00 value. PW3 leaves cycle stages and comparison bases unmatched. Do not convert the missing sample into zero or infer flock behavior or health from these synthetic traces.
- Interpreting a reset as negative consumption or a biological event.
- Filling an absent reading with zero or comparing cycle stages without a common basis.
Fields to preserve
- House and flock-cycle alias
- Defines the trace's scope.
- Cumulative reading, unit and reset evidence
- Supports interval reconstruction.
- Missing interval and cycle stage
- Preserves comparison limits.
Test another case
Remove the verified 510-liter pre-reset reading. Require A's full interval use to become unresolved rather than assume no use before reset.
Use fabricated traces reviewed by a qualified operator. Exclude integrator terms, actual telemetry and active animal-health records.
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.