For spatial overlay and analysis, raster data is better suited and more effective. This demanded the conversion of some vector layers.

Population density was classified into four categories (per square kilometer) (a) areas with a population density of 1500-3000 inhabitants , (b) areas with a population density between 3000-4000 inhabitants , (c) areas with a population density ranging from 4000- 5000 inhabitants , (d) areas with a population density <5000 inhabitants.

For weight allocation process, I used the Analytical Hierarchy Process (AHP) which involves pairwise comparisons of criteria. For the comparisons, a linear scale from 1 (which denotes a lack of preference) to 9 (which denotes a strong preference) is employed. A Consistency Ratio (CR) is computed in order to evaluate the consistency of the chosen preferences (Saaty, 1990), .

Forecasting method is highly subjective, even if it is restricted to the selection of the "best" forecasting method, the analyst's subjective judgment may have an impact on predicting (Baumann et al. 1998). I understand that the project is limited in its scope and can be further improved to achieve more accurate results by including more factors like water supply network, water consumption data, and present water pumping stations and ensuring inclusion of expert opinion.