SampleNext — Choose the next measurement without weakening the evidence

Collect more useful evidence

The next measurement is a decision. SampleNext explores how scarce, added laboratory capacity could produce more decision-relevant evidence without changing the scheduled monitoring that makes long-term trends trustworthy.

One observing system · four jobs

Better evidence is the goal.

Each lane answers a different question. Mixing them would make a targeted program look representative when it is not.

Required monitoring

The scheduled backbone continues unchanged so trends and regional coverage remain comparable.

Always fixed

Event-directed

Use an added assay where current observations suggest a possible harmful event.

Shown below

Learning-directed

Measure where the answer would reduce uncertainty and improve a consequential future choice.

Design requirement

Random check

Reserve a probability-selected share so low-ranked misses remain visible and the strategy can be judged honestly.

Design requirement

Event-directed demonstration

Try the transparent baseline

The combined historical analysis showed a promising signal: the available measurements were associated with elevated domoic acid. In a stricter retrospective replay—ranking samples month by month under a fixed 10% laboratory budget pooled across seven piers—both learned approaches captured more elevated samples than the simple count rule. But the advantage did not clear the uncertainty test set before the replay across both model types. It remains a lead to test, and more data from the real workflow are needed to determine whether it holds.

Synthetic demonstration · no recommendation

Sample details

Sample label *

Sampling date *

Location *

Microscope counts — cells/L *

Delicatissima group (blank if not reported)

Seriata group (blank if not reported)

Enter at least one group. A blank group is treated as unreported—not zero.

Reported-group sum—

Sea-surface temperature — °C * Required context. It does not change the count-based queue rank.

Chlorophyll — mg/m³ (optional) Optional—leave it blank when it is not available for this decision.

Current result

Ready for an example.Use invented values on the left. The tool will show how the transparent ordering rule behaves inside one example batch.

Synthetic demonstration

Added to the example batch

—

Ordering ruleHighest reported-group sum first

Required monitoringUnchanged

Artificial-intelligence modelNot used

Random-check laneDesign requirement; not automated here

This illustrates queue mechanics. It is not a toxin test, safety determination, or recommendation.

Transparent baseline queue

Add samples from the same decision batch. The highest reported-group sum ranks first; artificial intelligence is not choosing this order.

Added event-directed slots

Rank	Sample	Location	Reported-group sum	SST	Chlorophyll	Queue
No samples yet. Your first assessment will appear here.

Use invented values only. They stay in this browser tab; SampleNext does not upload or store them. This demonstration does not alter any monitoring schedule or make a safety decision.