Investigating water and energy consumer behaviour using ICT as part of Internet of Things
The hypothesis for this research is that when given meaningful information, consumers will be mindful of their water consumption. The project focuses specifically on water consumers in schools of the Western Cape in Cape Town, South Africa.
Current water bills are not a complete representation of consumer’s usage; one cannot pinpoint activities that are wasteful. These water bills are often only available to schools one or two months late. This makes it difficult for consumers to identify areas or times of water wastage.
The first goal of the study is to look at the complexities of human information processing. This will include a social-psychological view of water conservation behavior. The different aspects of behavioral change will be studied with a focus on the barriers of such change. Ways of including social and behavioral sciences in current conservation programs is one of the end goals of this research study.
The third goal is to obtain the usage data from consumers at schools in the area of study. Smart water meters and will be used to collect the data in near real-time. A randomised control trial (RCT) will be conducted on over 100 schools in the study area. The treatment of the RCT will be water usage information in the form of a report of each school's weekly water consumption and a second report that includes a competitive element among the schools. As part of the RCT, each school will receive basic water saving maintenance. The difference in difference method will be applied to the obtained data to determine the exact effect of the intervention and to quantify water use reduction.
The effects of the maintenance done at the schools will also be studied in detail, which the goal of identifying basic, affordable "quick fixes" that schools and the department of education can use to help schools save water and money.
The third objective of the study is to employ statistical learning methods to model and predict water consumption in the schools using the obtained historical data. The methods that will be used include ARIMA, Kth Nearst Neighbour regression and Kernel Density Estimation regression.
The study as a whole hopes to assist decision makers in the water management department. Firstly by predicting the amount of water that schools in the study area will use and identifying basic, affordable maintenance that the education department can conduct to save water and hence help schools redirect finances to the school's academic needs.