Real monitoring pipelines rarely trust a single feed. ACLED logs reported conflict and political-violence events on land; ADS-B broadcasts self-reported aircraft position, altitude and callsign; AIS does the same for ships; and GDELT mines global news coverage for events and tone in near real time. Each stream is independently noisy and partial — fusion means aligning them in space and time so that corroborating signals reinforce each other while single-source noise stays low-confidence.
Widening or narrowing the fusion radius trades precision for recall, exactly like a real pipeline: too tight and you miss genuine corroboration between feeds with GPS jitter; too loose and unrelated events on opposite sides of a region get merged into false correlations.
A rotating 3D globe simulates four independent open-data streams — conflict-event reports, aircraft transponders, ship transponders and global news mentions — and fuses them into corroborated alerts whenever they line up in space and time.
Red pulses mark simulated ACLED conflict clusters, amber flashes are GDELT news mentions, and blue/cyan markers are ADS-B aircraft and AIS vessels on real-world routes. A beam and alert count appear only when a news flash coincides with a nearby aircraft or vessel — the essence of multi-source corroboration.
Widen or narrow the fusion radius to see how spatial tolerance trades precision for recall, adjust the event rate to simulate quieter or busier news cycles, and toggle the ADS-B and AIS layers on or off to see how fusion confidence depends on which feeds are actually available.
Real pipelines fusing ACLED, ADS-B, AIS and GDELT face exactly this trade-off: GPS jitter and reporting delay mean a fusion radius that is too tight misses true corroboration, while one that is too loose merges unrelated events into false alerts.