Lauren Hanna, Business Development & Analyst, L3Harris Commercial Aviation
Using Expert Knowledge Fused with Data to Shift from What to Why
Within the aviation industry, the important areas such as passenger journey management, customer relationship management, flight operation safety, efficiency, air frame maintenance and engineering provide vast amounts of data needing analysis to deliver value. We have more data in these areas than we can imagine, for example consider an A380, it has roughly 4000 parameters which are recorded every 1-2 seconds; so a 10 hour flight would equate to well over 70 million pieces of data, just from the standard recording. The same situation exists in our training environments with complex simulators replicating real-world and worse case scenarios supported by our expert pilot trainers to improve pilot competencies to handle unforeseen situations. Collecting such data from every airplane on every flight, from every simulator on every training session, brings up the question of ‘How do you make sense of that?’ because everyone is aware that there is value to be gained from a safety and training position.
Historically as an industry we have been very good at analysing the data to establish WHAT has occurred. However, it’s increasingly important for us to now use the data to determine WHY something has occurred, or is occurring. Sometimes the reasons why reside in textual reports from the pilots, or from our domain experts interpreting the situation from multiple data sources. Fusing different data sets along with the ability to quickly and effectively compare against a benchmark allows for anomalies to be identified. Highlighting contributing factors to these anomalies enables us to easily discover what conditions were prevalent when these anomalous events happen. By fusing domain expertise from within the pilot training community we can further label data with the reasons why anomalies are occurring and make data driven decisions to improve our training scenarios and the competencies of our pilots.
Moving forward the management and analysis of unstructured data fused with expert knowledge will be important to the support of the shift within the industry from the what, to the why.