Poor recreational water quality poses a risk of acute gastrointestinal illness (AGI) to beach goers engaging in water activities. Internationally, concentration of Escherichia coli (E. coli) is regularly monitored to make risk management recommendations about beach water quality conditions. Laboratory processing times results in a delay in the availability of water sample results. As such, decisions about whether to post a beach as safe for swimming are based on previous day data. In Canada, these decisions are made by local public health units. Beach E. coli concentrations are influenced by several environmental and climatic factors, which can result in changes to water quality conditions in a short period. A greater understanding of these factors could assist in timely decision-making processes. We examined water quality data collected by public health authorities during the recreational bathing season from beaches across Canada, including freshwater beaches in Toronto, Niagara Region, and Manitoba, and marine beaches in Vancouver, and linked environmental data from federal and provincial weather stations. The first objective aimed to determine the environmental predictors of water quality in each region, as indicated by E. coli concentration. The second objective used this information to develop predictive models using a novel bayesian network approach. Bayesian networks have potential for high predictive power, require few assumptions, can be easily updated, present results in an intuitive format, and could be more cost-effective for routine use by public health and environmental authorities. This presentation will discuss the environmental factors associated with recreational water quality in the Canadian setting and present the bayseian network predictive models. The results of this project are relevant to recreational beaches worldwide as we provide insight on both marine and freshwater settings and may be used to inform beach monitoring programs and reduce the burden of water-borne disease.