Avionic Instruments is an original equipment manufacturer of power conversion products for the commercial aerospace and defense marketplace. An important part of their business model is aftermarket support of their products installed on existing aircraft in service. Of additional importance is having data related to aircraft operation and usage in order for Avionic Instruments to support their product, and their customers properly.
Avionic Instruments needed to accurately predict failure rates of installed equipment on existing aircraft using aircraft operational data and data related to historical product failure rates.
Three data sources were imported into the Polyture data warehouse.
The Aerospace Database contained prevalent information on aircraft usage across multiple models. The historical data from the CRM system presented data associated with product delivery dates. These data sources were then combined with ERP data using Polyture dataflows.
The resulting dataset was then fed into Polyture’s automated machine learning, where various models were trained, tested, and ranked by performance. The best performing model was then deployed into a production data pipeline. The resulting forecast was then displayed on a dashboard.
The goal of the finished pipeline was to understand rates of product demand from Avionic Instruments’ customer base, benchmarked against incoming data from the aerospace database in order to understand and predict the failure rates of Avionic Instruments’ products.
With these predictive models, Avionic Instruments is now able to plan supportability to its customer base, as well as sales and production demand, further in advance and with high confidence.