Modeling Los Angeles Water Leaks Using Different Measures of Temperature

Los Angeles, California, is facing record drought conditions. As a result, there is interest in all things related to building and maintaining water capabilities. Leaks in the infrastructure can lead to costly losses of water resources. Accordingly, attention increasingly is being devoted to water leak management. Using data available through the City of Los Angeles’ open data movement, the number of leaks is analyzed in order to study both the impact of temperature and whether the number of leaks is decreasing over time. Three different approaches for modeling the number of leaks, including regression, time series, and neural networks, are compared.