The past few years have seen a growing interest in Open Data, a movement that encourages the sharing of public sector information for better governance, greater services and more vibrant economy. While data in government agencies are in abundance, their value and quality vary significantly. Without a structured approach, low-value and poor-quality information assets may be chosen, limiting the positive impact of Open Data. In recognizing the research gap, we propose a five-phase process of data selection, drawing insights from the domains of business architecture and information quality. The process is described and illustrated with an example of local government at the city level. It can be applied to other government agencies as they embark on Open Data initiatives.