In a world where numbers and science are often taken as the voice of truth and reason, Quantitative Devices (QDs) represent the epitome of policy driven by facts rather than hunches. Despite the scholarly interest in understanding the role of quantification in policy, the actual production of rankings, forecasts, indexes and other QDs has, to a great extent, been left unattended. While appendixes and technical notebooks offer an explanation of how these devices are produced, they exclude aspects of their making that are arbitrarily considered "mundane." It is in the everyday performances at research centres that the micropolitics of knowledge production, imaginaries, and frustrations merge. These are vital dimensions to understand the potential, limitations and ethical consequences of QDs. Using two participant observations as the starting point, this thesis offers a comprehensive critical analysis of the processes through which university-based research centres create QDs that represent the world. It addresses how researchers conceive quantitative data. It pays attention to the discourses of hope and expectation embedded in the devices. Finally, it considers the ethics of creating devices that cannot be replicated independently of their place of production. Two QDs were analysed: the Violence Early Warning System (ViEWS) and the Environmental Performance Index (EPI). At Uppsala University, researchers created ViEWS to forecast the probability of drought-driven conflicts within the next 100 years. The EPI, produced at the Yale Centre for Environmental Law and Policy, ranks the performance of countries' environmental policies. This thesis challenges existing claims within Science and Technology Studies and the Sociology of Quantification that QDs co-produce knowledge within their realms. I argue that these devices act as vehicles for sociotechnical infrastructures to be consolidated with little debate among policymakers, given their understanding as scientific and objective tools. Moreover, for an indicator to be incorporated within a QD, it needs to be deemed as relevant for those making the devices but also valuable enough to have been previously quantified by data providers. Even more, existing sociotechnical inequalities, power relations and epistemic injustices could impede disadvantaged communities' (e.g., in the Global South) ability to challenge metrics originated in centres in the Global North. This thesis, therefore, demonstrates how the future QDs propose is unilateral and does not acknowledge the myriad possibilities that might arise from a diversity of worldviews. In other words, they cast a future designed to fit under the current status quo. In sum, through two QDs focused on environmental-related, this thesis launches an inquiry into the elements that make up the imaginaries they propose following the everyday life of their producers. To achieve this, I discuss two core elements. First, the role of tacit knowledge and sociotechnical inequalities in reinforcing power relations between those with the means to quantify and those who might only accommodate proposed futures. Second, the dynamics between research centres and data providers in relation to what is quantified. By scrutinising mundanity, this work is a step forward in understanding the construction of sociotechnical imaginaries and infrastructures.