Description of Session
Typically during the price estimation stage for global health supply chain initiatives, often there are many unknown variables that are needed for the estimation of freight costs for various shipping legs and lanes. These unknowns often drive poor price estimates which can lead to budget overruns later in the supply chain process. As these overruns can be more costly at a later point, this can lead to less funding available for additional global health-related product procurement/distribution, reducing the overall effectiveness of global health initiatives including a reduction in the number of lives saved. To address this issue, PFSCM developed and implemented a predictive modeling initiative through the use of historical supply chain data in a data warehouse, Python programming language and various stats and visualization libraries from Anaconda package manager and the Integrated Development Environment, Jupyter Notebook, and reporting and visualization through Microsoft Power BI. Through a process of objective definition, data gathering and treatment, integrated data modeling and visualization, PFSCM developed a regularly updated, user friendly, predictive modeling tool for staff involved with freight cost estimation. This initiative has saved time, reduced estimation uncertainty and improved budgeting, leading to improved supply chain effectiveness and reach.
