Dangerous Events Tracker

DET (Dangerous Events Tracker) is a situational-awareness platform designed
to convert individual events into shared, evidence-based Situations and to
expose actionable Calls to Action. The system is intended for scenarios
where power is asymmetric, and information fragmentation prevents
collective action. It prioritises clarity, ethical design, and
transparency over engagement or monetisation.

Core principles

  • Hexagonal architecture — the domain is independent of frameworks, persistence, or presentation layers.
  • Domain-driven — Events, Situations, and Calls to Action are the core entities.
  • Embedded Neo4j — persistence uses explicit Cypher queries; no ORM.
  • Transparency and ethics — avoid central narrative control; maximise verifiable evidence.
  • Evolving design — the design document is the authoritative source; every architectural decision is logged there.

The domain model

Three entities carry the whole model:

  1. Event — a single reported occurrence, subjective or objective, with an id, a timestamp, an (optionally anonymised) actor, context, and references to evidence.
  2. Situation — an emergent pattern of multiple Events, formed when aggregation thresholds are met. A Situation can be draft, validated, or actionable.
  3. Call to Action — triggered by a Situation, never by a single Event; it carries a suggested action, its beneficiaries, and the risk it mitigates.

The graph is explicit and typed: Events CONTRIBUTE_TO a Situation; a
Situation TRIGGERS Calls to Action; Events INVOLVE Actors and
OCCURRED_AT Locations. No auto-mapping, no ORM — queries are explicit and
readable, so reasoning stays auditable.

Request flow

Individual events (a worker not paid on time, unsafe conditions, a broken
window, a traffic light stuck red) are ingested via CLI, REST, or batch
import; they are aggregated into Situations when thresholds are met; and
Situations trigger evidence-based Calls to Action — never the other way
around.

Ethical and security considerations

  • Actor anonymity is preserved unless explicit consent is given.
  • Data verification and integrity are mandatory.
  • The system avoids monetising user-contributed data.
  • All decisions and thresholds are auditable.