Mapping digital governance projects through complex networks

A crucial need for policy design is to guide the planning and the allocation of funds in a targeted way, especially in developing countries. This requirement is particularly relevant in the expanding field of E-government. In this paper, we present a preliminary analysis of a public database of Digital Governance projects, based on the combination of Complex Network models and Natural Language Processing techniques. We embed projects into a complex network, in which the strength of a connection binding two projects is determined by the semantic affinity between their titles. We use an unsupervised community detection algorithm to partition the network in sets of similar projects, and interpret the results to unveil hidden information in the database. Project ratings provided by an independent evaluation group are inspected in order to investigate the strategical value of the topics underlying the communities, laying the foundation for novel quantitative methods of policy design.