 This paper proposes a framework that uses social media data and traditional methods to accurately predict traffic patterns for last-mile delivery logistics. It employs deep learning techniques such as graph convolutional networks and long short-term memory neural networks, as well as data from sources like coins and social media check-ins. Additionally, it incorporates simulation tools such as agent-based simulation, discrete event simulation, and system dynamics to further enhance vehicle routing. This article was authored by Valeria Lanes-Viascunari, Edgar Gutierrez Franco, Luis Rebello, and others.