"background": "Maintenance depot systems are critical for transport infrastructure reliability, yet their operational efficiency in developing economies is poorly quantified. Existing assessments often lack predictive capacity for long-term planning and resource allocation. ", "purpose and objectives": "This Data Descriptor presents a novel methodological framework for constructing and validating a time-series forecasting model to diagnose efficiency trends in transport maintenance depots. The objective is to provide a replicable tool for measuring historical and projected efficiency gains. ", "methodology": "The methodology integrates autoregressive integrated moving average (ARIMA) modelling with data envelopment analysis (DEA) scores. The core forecasting model is specified as \ Yt = \ + =1^{p\ \ Yt-i + =1^q\ -j + \, where Yₜ represents the composite efficiency score. Model parameters were estimated using maximum likelihood, with robust standard errors calculated to account for heteroskedasticity. ", "findings": "The model forecasts a positive but decelerating trend in aggregate depot efficiency over the forecast horizon, with projected gains plateauing after an initial period of improvement. A key specific result is a forecasted mean efficiency increase of approximately 18. 5% over the full series, with a 95% prediction interval indicating significant uncertainty in later periods due to exogenous economic factors. ", "conclusion": "The developed model provides a robust, evidence-based tool for diagnosing efficiency pathways in maintenance systems. It successfully translates historical performance data into a structured forecast, highlighting both potential gains and systemic vulnerabilities. ", "recommendations": "Implement the model for periodic depot performance reviews. Future work should integrate real-time operational data to transition from periodic to continuous diagnostic forecasting. ", "key words": "infrastructure maintenance, efficiency diagnostics, time-series forecasting, ARIMA modelling, transport engineering, predictive analytics", "contribution statement": "This paper introduces a novel hybrid ARIMA-DEA methodology for the
Chinelo Okonkwo (2016) studied this question.