Publicação

Data Analytics and Machine Learning in the Aviation Industry Supply Chain

Ver documento

Detalhes bibliográficos
Resumo:Amid growing global volatility, the aviation industry's supply chain faces new challenges, like supplier unreliability, geopolitical disruptions, and unpredictable lead times for critical parts. These issues demand proactive approaches that go beyond traditional procurement analytics. This thesis examines how data analytics and machine learning can be leveraged to anticipate and mitigate such disruptions, utilizing real procurement data from a global aviation company. Can machine learning forecast the unexpected in the new aviation supply chain? Could a datadriven approach reshape how we manage the supply chain? These questions guide the development of an artifact that supports decision-making through machine learning. Following the CRISP-DM methodology, the research integrates unsupervised clustering techniques to segment supplier behavior and supervised regression models to predict part delivery lead times. The Random Forest model achieved strong predictive performance (R² = 0.92; MAE = 0.35 days), while agglomerative clustering revealed interpretable supplier groups that reflect a new way to manage supplier relationships in the supply chain. The results show that machine learning techniques can improve strategic supplier management and short-term operational planning in the aviation industry. This research contributes to the growing body of literature on data-driven supply chain optimization, providing a scalable foundation for future integration into business intelligence planning systems.
Autores principais:Barrela, Francisco Correia
Assunto:Machine Learning Supply Chain Aviation Data Analytics Regression Clustering SDG 4 - Quality education SDG 9 - Industry, innovation and infrastructure
Ano:2025
País:Portugal
Tipo de documento:dissertação de mestrado
Tipo de acesso:acesso aberto
Instituição associada:Universidade Nova de Lisboa
Idioma:inglês
Origem:Repositório Institucional da UNL
Descrição
Resumo:Amid growing global volatility, the aviation industry's supply chain faces new challenges, like supplier unreliability, geopolitical disruptions, and unpredictable lead times for critical parts. These issues demand proactive approaches that go beyond traditional procurement analytics. This thesis examines how data analytics and machine learning can be leveraged to anticipate and mitigate such disruptions, utilizing real procurement data from a global aviation company. Can machine learning forecast the unexpected in the new aviation supply chain? Could a datadriven approach reshape how we manage the supply chain? These questions guide the development of an artifact that supports decision-making through machine learning. Following the CRISP-DM methodology, the research integrates unsupervised clustering techniques to segment supplier behavior and supervised regression models to predict part delivery lead times. The Random Forest model achieved strong predictive performance (R² = 0.92; MAE = 0.35 days), while agglomerative clustering revealed interpretable supplier groups that reflect a new way to manage supplier relationships in the supply chain. The results show that machine learning techniques can improve strategic supplier management and short-term operational planning in the aviation industry. This research contributes to the growing body of literature on data-driven supply chain optimization, providing a scalable foundation for future integration into business intelligence planning systems.