{ "background": "Agricultural productivity in sub-Saharan Africa remains constrained by inefficiencies in cultivation and resource management. Process-control systems, integrating sensor networks and automated actuators, present a potential engineering solution, but rigorous methodological frameworks for evaluating their causal impact on crop yield are lacking.", "purpose and objectives": "This Data Descriptor presents a structured dataset and methodological framework designed to quantify the causal effect of adopting agricultural process-control systems on maize yield. The primary objective is to provide a replicable difference-in-differences (DiD) data architecture for impact evaluation in engineering applications.", "methodology": "The dataset comprises panel data from smallholder farms, including treatment units with installed control systems and control units without. Core variables are plot-level maize yield (tonnes/hectare), a binary treatment indicator, and time-period dummies. The causal impact is estimated using a two-way fixed effects DiD model: $Y{it} = \\beta0 + \\beta1 (\\text{Treat}i \\times \\text{Post}t) + \\alphai + \\gammat + \\epsilon{it}$, where robust standard errors are clustered at the farm level.", "findings": "As a Data Descriptor, this paper presents the dataset structure and methodological framework, not empirical results. The constructed dataset is designed to detect a minimum detectable effect of a 15% yield improvement with 80% statistical power. Preliminary data validation indicates that the parallel trends assumption, critical for DiD, is testable with the provided pre-intervention data.", "conclusion": "The described dataset and DiD framework provide a robust foundation for isolating the causal impact of engineering interventions in agriculture, moving beyond correlational analysis.", "recommendations": "Researchers should utilise the provided diagnostic variables to test the parallel trends assumption before applying the DiD model. Future data collection should incorporate more frequent temporal data points to analyse effect dynamics.", "key words": "causal inference, precision agriculture, panel data, fixed effects, impact evaluation, sensor