The FAA’s Data Science University Engagement program, hosted by the Chief Data Office, is a partnership between the FAA and academia. The program is providing graduate-level students with the opportunity to solve real-world data science challenges and is producing tangible results for the agency.

Read about two recent projects with George Mason University (GMU) to learn more about this amazing opportunity for partnership.

Determining Noise and Actuals in the Sighting/Incidence Database

This GMU team tackled a project, co-sponsored by APO-100 and AFS-930, that challenged students to leverage their data analysis expertise to validate UAS data reports and observations. At the end of the project, the students delivered a script optimized to effectively clean and extract vital information from the FAA’s UAS sighting reports to merge with external data sources. This new workflow will aid the FAA in establishing a streamlined process for taking raw reports and turning them into a data source that allows for detailed analysis of UAS incidents through visualizations and summary statistics. Read more here.

Applying NLP to Analyze Large Volumes of Safety Event Reports

Another GMU team provided a jump-start for the FAA’s Air Traffic Safety Oversight Service (AOV) safety research initiatives by developing a model to automate the process of extracting data from text reports using machine learning techniques. The team applied natural language processing (NLP) to identify hazard information contained in events reported in the Administrator’s Daily Alert Bulletin (ADAB). The goal was to identify key pieces of information from each narrative, based on an over-arching hazard classification system, so that information gleaned from the ADAB can be related to similar information from other sources. Read more here.