The field of Computer Science has seen increasing interest in off-grid communities systems due to their importance for sustainable development in South Africa and other resource-limited regions. The study will employ a systematic literature review approach, synthesizing existing research from various databases and journals. Methodological rigor will be assessed using predefined criteria, including the use of randomization techniques for experimental design. A key finding is that many studies lack robust methods to ensure system reliability in off-grid communities, particularly concerning data collection strategies and statistical models used to measure outcomes. The identified methodological weaknesses underscore the need for a randomized field trial to provide reliable evidence on system performance under varying conditions. Recommendations include integrating best practices from existing studies into future trials, such as using validated statistical models and ensuring transparent data collection protocols. Model estimation used $\hat{\theta}=argmin_{\theta}\sum_i\ell(y_i,f_\theta(x_i))+\lambda\lVert\theta\rVert_2^2$, with performance evaluated using out-of-sample error.