Font Size: 

In the literature on topic modelling, network-based procedures for topic detection have become popular as an alternative to classical topic models, showing promising results. However, the lack of a systematic analysis of how the design choices made in text processing and network definition affect the results in terms of topics detected makes using these procedures demanding. Therefore, this work aims to fill this gap by showing how and to what extent the choices made during the analysis influence the features of the topics discovered.