Boeing Insight Accelerator – Webinar Responses

November marked a significant milestone for Insight Accelerator as we completed our first broad public outreach through a set of webinars for Boeing airline customers. We were fortunate to engage with representatives from dozens of airlines, including both flight ops and tech ops teams, on the following critical issues facing their maintenance organizations:

 

Developing good predictive alerts

 

Key customer issues:

  • Lack of appropriate big data infrastructure and/or data processing pipeline
  • Lack of data science skills and/or access to qualified resources

As a cloud-based Software-as-a-Service (SaaS) application, Insight Accelerator provides an integrated, scalable environment for exploring, processing and monitoring flight and maintenance data that is focused on empowering engineers to use analytics without specialized training in programming and data science. Airlines no longer need to build and maintain on-premises big data environments to facilitate predictive analytics, nor do they need to employ data scientists to take advantage of the latest machine learning technology.

 

Operationalizing predictive alerts

 

Key customer issues:

  • Too many false alerts
  • No efficient process for alert development and maturation

The ability to characterize the performance of a predictive alert using historical data and before deploying the alert into live operations is a key feature of Insight Accelerator. Airlines can tune the algorithm thresholds and tailor alerting criteria to suit their sensitivity to false alerts, leading to a better understanding of the expected accuracy performance of each algorithm. Insight Accelerator includes a built-in process that guides the user through the steps of data exploration, data feature evaluation, algorithm testing and deployment to bring mature, well characterized alerting algorithms into an operation.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.