Trace — Shipping Investigation & Fraud Detection
Giving investigators a navigable picture of delivery anomalies
When unusual shipping behavior spans locations and events, an investigation needs more than a table of package IDs. It needs controlled access, search and geography working together.
The challenge
Trace is a logistics investigation platform for examining anomalous delivery patterns. It brings geospatial package search into a map-based workflow.
The documented stack includes React, Spring Boot and Elasticsearch, with OAuth2/OIDC role-based access control and a WebGL-accelerated map interface.
System approach
Search by spatial context
Geospatial package search lets an investigator reason about where events happened, rather than depend solely on individual tracking identifiers.
Make the map an investigative interface
The WebGL-accelerated view supports exploring geographically distributed patterns, giving the map a functional role in reviewing anomalies.
Control access by role
OAuth2/OIDC RBAC is part of the documented implementation. Investigation data requires clear authorization boundaries across the workflow.
Engineering considerations
These are design principles the work raises, rather than a claim about private architecture or unreported implementation choices.
Map performance follows data density
For similar systems, bounding the queried geographic area and aggregating dense markers can keep interaction useful. Those techniques are illustrative considerations, not a claim that Trace uses a particular tiling strategy.
Search permissions must survive every layer
Protecting the UI alone is insufficient: query and API authorization should enforce the same access model.
Help people distinguish signal from noise
An anomaly is a starting point for investigation, not proof of fraud. Interfaces should preserve enough context for a person to make that distinction.
Scope and outcome
The published capabilities include geospatial search, a WebGL map UI and role-based access control. Investigation throughput, fraud detection rates and other private operating results are not asserted.
This account describes the project at a public level and labels broader design considerations separately from implemented capabilities.