Monitoring the Water Utility Performance in Drinking Water Quality Compliance using Data Mining Approaches

Monitoring the quality of water ensures it is safe for life. Nowadays, its effectiveness requires embracing data science. This was done in the context of the Water and Sanitation Corporation (WASAC) of Rwanda, which manages 18 water treatment plants (WTPs). This research aims at monitoring the drinking water quality key performance indicators (KPIs) using an interactive dashboard that can map interventions, perform calculations, assess the performance, and provide near real-time information. A four-step approach—KPI identification, data collection, monitoring tools, and progress tracking—was used. Step 1 used the literature review to identify KPIs in terms of compliance with clean water regulatory and reliability requirements, and the existence of a functional system to address customer needs. Then, primary and secondary data sets, including Twitter data, were collected in step 2. The third step consisted of computing the drinking water quality index (DWQI), performance discrepancy scores (PDS), and Twitter sentiment scores, as well as an interactive dashboard. Finally, results were analyzed to determine WASAC's performance towards quality KPIs. Compliance with clean water requirements has been generally consistent, and the first five WTPs were Kadahokwa, Cyunyu, Mutobo, Muhazi, and Kimisagara. However, the water reliability could not be confirmed due to missing data. Twitter data analysis confirmed the existence of a system to address customer needs and to notify customers of planned rationing or interruptions of water supply. Therefore, the suitability of water for human consumption was generally established, but more efforts were recommended to ensure full and consistent compliance.