Solar radiation, the electromagnetic energy emitted from the Sun, is a fundamental driver of Earth's weather and climate systems and represents a vast, clean energy source. Global solar radiation (GSR) is the total amount of solar radiation (both direct and diffuse) reaching a horizontal surface on Earth. It is a key measurement for evaluating the solar energy potential of a specific location, which in turn is essential for assessing the expected performance and efficiency of solar photovoltaic (PV) system. Thus, the accurate measurement and estimation of GSR is crucial for assessing and utilizing solar energy resources at both global and local scales. Hence, this study investigates estimations of GSR and assesses the performance of crystalline silicon (c-Si) PV cells/modules across Ethiopia, by utilizing a comprehensive approach that integrates data-driven and physical models. The study also implemented various optimization techniques such as determining the optimum tilt angle and tracking mechanisms. For this purpose, twelve machine learning (ML) and one stacked/ensembled model were trained and validated with hourly, daily and monthly ground-based global solar radiation data from 16 synoptic weather stations (2020-2022), supplemented by meteorological, aerosol, and sky condition data from MERRA-2 and NASA POWER archives. The three stations with distinct weather patterns were withheld from the model development process for model transferability/generality test. A stacked/ensemble model (i.e., constructed by stacking better performing separate models) showed exceptional predictive performance with error metric values ranging (R²: 0.956-0.963; RMSE: 9.938-11.784 W/m²) for all time scales. With this performance capability we generated a high-resolution (1° x 1°) global solar radiation data across Ethiopia for the year 2022, and the distribution showed a precise spatial and seasonal dependence with the highest in spring (i.e., 594 - 641 W/m2; eastern and northeastern) and lowest in summer (i.e., 359 – 405 W/m2; western and southern parts of the nation). Such analogs were also observed on the peak sun hours and plane-of-array (POA) irradiance distribution with their annual value ranging from 4.83 – 6.57 kWh/m2/day and 0.65 – 1.05 kW/m2, respectively, across the nation. Here it’s worth noting that to model POA irradiance, we implemented five decomposition and six transposition models (i.e., thirty different independent combinations). Furthermore, we incorporated POA irradiance into a single diode PV cell model to evaluate c-Si PV cell performances. Consequently, the annual PV cell temperature, ranging vii from 38.84°C to 55.45°C, significantly impacted device parameters. The short-circuit current (Jsc) mirrored POA irradiance trends, while the open-circuit voltage (Voc) showed an inverse temperature dependence. The annual PV cell efficiency varied from 13.02% to 22.35%, with clear average seasonal variations such as the highest in spring (20.72%) and the lowest in summer (16.01%). Besides, optimal PV module tilt angles and implementation of different tracking mechanisms were determined. The monthly optimal tilt angles ranged from 0° (July) to 47.90° (January), while seasonal averages were observed as 29.40° (winter), 21.65° (autumn), 12.34° (spring), and 8.8° (summer). The annual optimal tilt angle varied from 14.51-21.52o. In addition, the performance of different tracking mechanisms (dual/full axis: DAT, vertical-axis: V-axis, east-west/incline east-west: EW/IEW, north-south: NS) were also evaluated. Dual/full axis tracking yielded the highest annual average efficiency (44.89%), while NS tracking resulted in a 28.46% energy loss compared to horizontal mounting. On the other hand, PV module mounted at optimal tilt angle resulted in a 4.12% gain, while EW/IEW and vertical axis tracking yielded 43.12% and 24.94% energy gains, respectively compared to horizontally mounted. Overall, this study provides a valuable resource for selecting, comparing, and designing future solar PV systems in Ethiopia. It contributes to strategic PV deployment by aiding in energy assessment and forecasting. Moreover, the study offers an optimal tilt angle model, guiding the selection and implementation of the best tracking mechanisms for PV modules/panels in Ethiopia. This information is essential for effective energy assessment and forecasting across the nation