Advances in Internet technologies have made it possible to gather, store, and process large quantities of data, often in real time. When considering smart and sustainable cities, this big data generates useful information and insights to citizens, service providers, and policy makers. Transforming this data into knowledge allows for empowering citizens' cognition as well as supporting decision-making routines. However, several operational and computing issues need to be taken into account: 1) efficient data description and visualization, 2) forecasting citizens behavior, and 3) supporting decision making with intelligent algorithms. This paper identifies several challenges associated with the use of data analytics in smart sustainable cities and proposes the use of hybrid simulation-optimization and machine learning algorithms as an effective approach to empower citizens' cognition and decision making in such ecosystems.