The Internet poses a number of challenges for academics. Internet specificities such as anonymity, the decontextualisation of discourse, the misuse or non-use of references raise methodological questions about the quality and the authenticity of the data available online. This is particularly true when dealing with extremist groups and grass-root militants that cultivate secrecy. Based on a study of the far-right on Twitter, this article explores these methodological issues. It discusses the qualitative indicators we have developed to determine whether a given Twitter account should be included in the sample or not. By using digital traces drawn from profiles, interactions, content and through other visual information, we recontextualize user’s profile and analyze how digital traces participate in providing far right ideas with a wider representation.