I recently have been analyzing historical flight data. This includes an analysis of reductions in flight operations during the COVID-19 pandemic and its impact on aviation noise for communities in the United State (preprint). While that analysis utilized data sourced from FAA databases, a more flexible and complete flight data are available from crowdsourced logging of ADS-B flight radio signals across the world.
Logging of ADS-B signals is possible using low-cost radios. I have my own <$100 software defined radio (SDR) setup that I can use to record signals from flights in my area. This is useful for real-time information on overhead flights.
For historical data, I am currently using ADS-B flight data from OpenSky. These data are available for many parts of the world. In order to access historical OpenSky data, there are recommended R and Python software packages that facilitate querying their online Trino database. As an R user, though I found the recommended openSkies R package difficult for my use case.
I want to download ADS-B flight data from the Washington State area for all of the year 2025. The openSkies R package by Rafael Ayala and colleagues is excellent, in that it provides various “get” functions for querying the historical database. However, because they wrote their package using object-oriented r6 classes, the downloaded data are extremely difficult to serialize (i.e., difficult to save to disk), taking a huge amount of time and space to save because saving r6 objects includes not only data, but functions and the R environment.
I forked the openSkies R package, and changed the “get” functions to include an asDF parameter, which when asDF=TRUE, allows the query resulting data to be returned as an R data frame, instead of the r6 objects. This makes saving the data take fractions of a second instead of many minutes — and the data in data frames is tiny compared to r6 objects.
You can install my forked version of the openSkies R package from my github site: https://github.com/edmundseto/openSkies