Evidence Explorer¶
Maturity: PUBLIC_ALPHA · local read-only research surface
Evidence Explorer is the local NosoGraph interface for inspecting a typed claim, its evidence direction, study or source context, and available provenance. This page is procedural: it explains what to open and how to interpret the records without turning an evidence label into a scientific conclusion.
Research boundary. Evidence Explorer presents computational research artifacts.
SUPPORTSdoes not mean proven,CONTRADICTSdoes not by itself establish falsification, and an association is not causation. NosoGraph is not medical advice or clinical decision support.
Open a claim¶
- Start the local dashboard from the Docker guide or the API/server guide.
- Open the dashboard at http://127.0.0.1:8000/.
- Open a condition, choose a claim, and select Open in Evidence Explorer. Depending on the local dashboard build, the same surface may be reached from the Evidence navigation item.
- Inspect the claim row, evidence groups, source identifiers, and provenance details.
A shareable claim URL uses the documented shape ?claim_id={uuid}#evidence-explorer. The UUID is an application identifier, not a scientific confidence value.
Read the claim¶
A claim is a structured statement with a subject, predicate, and object. The predicate names the relationship; it does not decide whether the relationship is causal or clinically useful. The Claims concept defines the model and the Evidence model explains how records attach to it.
Read evidence direction¶
The canonical evidence directions are:
| Direction | Meaning in NosoGraph |
|---|---|
SUPPORTS |
The evidence record is represented as supporting the claim. |
CONTRADICTS |
The evidence record is represented as disagreeing with the claim. |
INCONCLUSIVE |
The available evidence is mixed or does not establish a clear direction. |
UNASSERTED |
No directional evidence is recorded for the claim. |
These labels describe the stored relationship between an evidence record and a claim. They are not a universal confidence score, proof state, or clinical recommendation. Supporting and contradictory records can coexist; inspect each group separately. See the canonical Evidence semantics page for the full interpretation reference.
Read context and missingness¶
Quality dimensions such as species, study design, origin, and human review describe context. A field that is not present should remain explicitly unavailable; it should not be silently converted into a negative finding or a score.
In Compare outputs, KNOWN_ABSENT means an explicit current negated assertion exists, while NOT_RECORDED means neither a positive nor a negated assertion is present. Those are different contracts; the Compare guide defines them in detail.
Trace provenance and sources¶
Follow the available chain:
claim -> evidence -> study or source -> snapshot/context -> provenance
Source identity, import context, version/date, fingerprints, and incomplete stages are shown only where the current record supports them. Provenance establishes traceability to stored inputs; it does not establish that the underlying conclusion is correct. See Provenance and the source matrix.
API view¶
A loaded claim can expose its structured response through the local API. The evidence list is paginated:
GET /api/v1/claims/{claim_id}/evidence?limit=50&offset=0
Use the API reference for the current base URL and endpoint catalog. The route is local unless you deploy the application yourself.
Continue¶
- Claims — understand the subject, predicate, and object.
- Evidence semantics — use one canonical direction vocabulary.
- Trace evidence — follow the inspection sequence.
- Provenance — interpret source and snapshot metadata.
- Compare conditions — inspect recorded differences without implying diagnosis.