Water is a vital resource in practically every aspect of our lives, including our homes, where water is used for cleaning, drinking, and a variety of other activities. Governments have already put in place strategic measures to supply this natural resource in urban and rural areas where people dwell. However, due to poor administration and lack of monitoring, millions of people continue to live without full access to these resources. Inadequate management of water resource services after they are produced and given, on the other hand, can lead to poor social economic development and human well-being. Consequently, this can lead to water planning or inability to the implementation of the strategic plans that the governments have already set. The decision-making about water planning should also focus on the population growth of the targeted region.

In the current study, we developed not only an IoT-based intelligent system for monitoring and managing household water use, but also a predictive machine learning model for predicting household water consumption based on population increase in the targeted region. We created an IoT smart meter that can track household use and is connected to a server in the XAMP platform to monitor household water usage.

For machine learning modeling, we have built a predictive model in python using data merged from different institutions. Mainly 7 years water data from WASAC (2014-2021) and population data from Kigali city extracted based on two last consecutive censuses (2002-2012). The predictive model was built using Random Forest algorithm and gave 96.1% and 91.3% of training and testing accuracies respectively. And finally the IoT based intelligent household consumption has been prototyped to monitor water consumption via dashboards by installing IoT smart water meter at household.