We present NETSCITY, an online application to analyse and visualise world scale scientific production and collaboration data between cities. Contrary to existing tools that mainly focus on displaying co-occurrence networks, NETSCITY especially focuses on processing the geographical information comprised in bibliometric data. NETSCITY proposes a fully integrated solution to parse and clean the authors' addresses, comprised in a set of references, geocoding them at the city level, clustering them at the requested level of analysis (urban areas or countries) and mapping them either on a world base map or in a relational space. In the first part of the paper, we stress the originality and design of the NETSCITY application in terms of geocoding, clustering, counting, and mapping methods. In the second part, we detail its main functionality as a geoweb application. Eventually, we show the results one can obtain by applying NETSCITY on a set of references extracted from the online version of the Web of Science Core Collection.