Existing studies in Developing and Newly-Industrialised (DNI) countries merely analysed the load demand forecasting methodologies, with little emphasis on consideration of data quality, provision of research agenda and the proposition of the design architecture of load forecasting. Therefore, this paper surveys 22 articles from 18 DNI countries, attempting to investigate the load forecasting methodologies as well as data cleansing mechanisms, and then proposes a general design framework for these countries. A systematic review protocol is applied in this study to achieve unbiased and scientific-based findings. Economic growth, number of customers, price of electricity, temperature, calendar events, and daytime found to be significant drivers of electricity consumption. Furthermore, the findings indicate that 63.64% of the surveyed load forecasting mechanisms considered the inclusion of outlier-removal preprocessors. This paper has pinpointed the issues pertaining to load forecasting in the DNI countries such that a robust model, fitting a context can be built with efficiency.