A worldwide, machine-generated airfield database: better than hand-curated datasets?

Throughout the last decade, the rise of open data in research has led to the release of many scientific datasets. Air transportation data are of particular interest given their high impact on migration, epidemic models and so on. Existing datasets are often specifically hand-curated, with the support of volunteers. One major problem with such datasets is that the quality, in terms of correctness and coverage, of the data is often not clear. Commercial datasets, on the other hand, are expensive and unnecessarily increase the entrance barrier for new researchers. In this work, a worldwide, machine-generated airfield database, which includes information on airports, helipads and runways serving airports, was derived. The database was generated using three machine-processable open datasets as input (OpenStreetMap, DBPedia and Wikipedia). Advanced data management techniques enabled the generation of an airfield database comparable with the best existing commercial, hand-curated solutions. This work contributes towards using open big data in air transportation. Furthermore, this machine-generated database, which can be automatically updated at any time, will hopefully raise the interest of other researchers in the community and lay the foundation for a gold standard for air transportation research.