An integrated data analysis and machine learning approach to track and monitor SARS-CoV-2 in wastewater treatment plants

Wastewater-based epidemiology (WBE) programs are cost-effective for continuously monitoring infections, including SARS-CoV-2. This work proposes combining data analysis and machine learning to track and monitor SARS-CoV-2 in wastewater treatment plants. Our approach includes exploratory data analysis and data regression using support vector machine regression (SVM) models fitted with the collected data by New York City (NYC). SVM regression models show a coefficient of correlation R 2 between 0.93 and 0.99 compared with linear regression models reporting values within 0.70–0.88. Moreover, we propose the estimation of the optimal sample size using Monte Carlo analysis and the corresponding operational cost reduction of the existing NYC program as a result of this optimization, estimated as 170,000 USD per year (~ 40% decrease). Our approach can be used to optimize existing and new WBE programs. Thus, we run a quick economic exercise for a case study in the Global South to provide a clear picture of the capital expenditure (CAPEX) and operational expenditure (OPEX) breakdown structure for implementing these programs.