← All projects

StageCompletedAreaData and analysis

Flight paths and profiles from ADS-B data

Analysis of departure and arrival paths and profiles at Warsaw Chopin Airport, prompted by questions about aircraft noise.

Aircraft departure trajectories and climb profiles.

Problem

Watching aircraft over the area made me curious about their routes and how those relate to noise. Collecting flight data was only the first step; understanding the patterns required analysis.

Solution

What began as manual analysis of ADS-B data grew into a project examining flight paths and altitude profiles. I prepared a report on aircraft departing Warsaw Chopin Airport. I also analyse approach and climb profiles in the context of CDA, NADP1 and NADP2.

My contribution

Building the analysis and writing a report on departures from Warsaw Chopin Airport.

From collecting data to investigating flight patterns

I have been receiving ADS-B signals and sharing the data with flight tracking services for years. Questions about aircraft noise led me to look more closely at routes. Manual checks grew into a broader analysis of flight paths and altitude profiles during departures and arrivals.

The report and ongoing work

I prepared an analytical report on departures from Warsaw Chopin Airport. The HTML version is linked in the materials below; the report continues to be developed and updated.

What I examine in the data

Report v11 brings together two analyses: a yearly snapshot of SID departure routes (19 May 2025 – 19 May 2026) and a comparison of climb profiles in two 48-day windows (19 June – 5 August and 8 August – 24 September 2026).

The NADP dataset covers 29,721 departures. The analysis compares profile similarity to NADP1 and NADP2, includes unresolved classifications and checks sensitivity to QNH altitude correction.

The night-time increase in NADP1-like profiles is significant in the baseline variant, but does not remain significant after QNH correction and adjustment for multiple comparisons. The report does not present this as robust evidence of either change or no change.

Limits of interpretation

Trajectories show flight geometry, not the content of air traffic control instructions. Profile classes are proxies; they do not confirm that a particular procedure was followed or that rules were breached. ADS-B data alone is not a noise measurement.

Version v11, dated 12 October 2026, is a labelled snapshot: it contains no October flights, and the historical SID section was not recalculated in this edition. The conclusions apply to the recorded sample; the receiver does not capture every flight or every part of a flight. Data, denominators and limitations are described in the linked report.

Technologies

ADS-B, Raspberry Pi, dump1090, Python, FastAPI, DuckDB, Shapely, Matplotlib

Links and resources

Last updated: