AeroScope began with a simple question — "what is flying over me, and can I trust it?" — and a conviction that the answer should be transparent. It is built and maintained by one independent researcher, runs on open data, and refuses to dress up rules as magic.
Dozens of volunteer networks already publish the world's ADS-B traffic for free, but as undifferentiated dots. AeroScope's job is to turn that firehose into answers about a specific patch of sky: what each aircraft is, how it is behaving, whether its signal can be trusted, and which few deserve a second look. Everything is observer-relative — anchored to a point you choose, not to a named flight.
The detection stack is deliberately torch-free — River, scikit-learn, pykalman, OpenAP, stumpy, PyOD — chosen because every flag is auditable. There is no PyTorch and no reinforcement learning; an earlier RL prototype was evaluated and retired.
The ADS-B Anomaly Benchmark is published CC-BY 4.0 with documented columns, a baseline result and honest limitations — so the work can be checked, not just believed.
AeroScope ranks attention; it does not judge intent. It sees only what aircraft broadcast. Where a capability has edges — silent drones, EMCON aircraft, sparse fields — the platform says so plainly.
Independent researcher, Department of Computer Science, COMSATS University Islamabad. AeroScope is already live, already open, and already shipping citable data. If you are a researcher, a receiver-network operator, an investor or an organisation that cares about ADS-B integrity and open aviation data, I'd genuinely like to talk.