This paper argues that visualisation conventions work to make the data represented within visualisations seem objective, that is, transparent and factual. Interrogating the work that visualisation conventions do helps us to make sense of the apparent contradiction between criticisms of visualisations as doing persuasive work and visualisation designers’ belief that through visualisation, it is possible to ‘do good with data’ (Periscopic, 2014). We focus on four conventions which imbue visualisations with a sense of objectivity, transparency and facticity. These include: a) two-dimensional viewpoints; b) clean layouts; c) geometric shapes and lines; d) the inclusion of data sources. We argue that thinking about visualisations from a social semiotic standpoint, as we do in this paper by bringing together what visualisation designers say about their intentions with a semiotic analysis of the visualisations they produce, advances understanding of the ways that data visualisations come into being, how they are imbued with particular qualities and how power operates in and through them. Thus this paper contributes nuanced understanding of data visualisations and their production, by uncovering the ways in which power is at work within them. In turn, it advances debate about data in society and the emerging field of data studies.