Multivariate time series is a concept common in the real world. For example, many financial assets show dependence on each other and macroeconomic factors like interest rates, inflation, and non-economic factors like weather. This paper describes a computationally efficient approach to clustering and forecasting these multivariate time series.