{ "background": "Transport maintenance depots are critical infrastructure for economic development, yet systematic evaluation of their operational yield in West Africa remains underdeveloped. Current assessments often lack robust statistical frameworks to account for hierarchical data structures and inherent uncertainties.", "purpose and objectives": "This study aims to develop and validate a Bayesian hierarchical model to quantify and analyse yield improvement within transport maintenance depot systems. The objective is to provide a methodological framework for identifying key performance drivers and predicting system efficiency.", "methodology": "We formulate a Bayesian hierarchical model where depot-level yield $Y{ij} \\sim \\text{Normal}(\\mu{ij}, \\sigma^2)$, with $\\mu{ij} = \\alpha + \\beta X{ij} + uj$, and $uj \\sim \\text{Normal}(0, \\tau^2)$. The model incorporates depot-specific random effects $uj$ and covariates $X{ij}$ on spare parts inventory and technician training. Inference is based on Markov chain Monte Carlo simulation using data from a sample of depots.", "findings": "The model identified a positive association between technician training hours and yield, with a posterior mean for the coefficient $\\beta$ of 0.15 (95% credible interval: 0.08 to 0.22). This indicates that a 10-hour increase in training is associated with a 1.5-unit yield improvement. The depot variance $\\tau^2$ was estimated with substantial uncertainty, highlighting system heterogeneity.", "conclusion": "The proposed Bayesian hierarchical model offers a statistically rigorous tool for evaluating maintenance depot performance, effectively handling multi-level data and quantifying uncertainty. It demonstrates that targeted investments in human capital are a significant lever for yield improvement.", "recommendations": "Depot managers should prioritise structured technician training programmes. Policymakers are advised to adopt this modelling framework for systematic performance benchmarking and resource allocation across the national depot network.", "key words": "Bayesian inference, hierarchical modelling, maintenance engineering, infrastructure