This paper reports on a data sprint conducted as part of a PhD course on digital methods and data critique at the [University name redacted for peer review]. We reflect on how our data sprint contributed to this higher educational setting, and point to ways in which the data sprint method can be developed further based on our experience. The paper discusses how the sprint fabricated a moment of “critical proximity” for students that were mainly working with qualitative social science methods. The data sprint allowed them to put their critique on “big data” into practice by working with selected sets of data from Twitter and Scopus. We reflect on our collective experience and draw conclusions on the use of data sprints in teaching. Data sprints encourage us to engage with feelings of being underwhelmed and overwhelmed by data that provoke our social science way of critique. Our data sprint tangibly demonstrates that data work is in fact “messy”: transgressing ideals of good data management, biassed, ambiguous and open-ended. But instead of turning away from this “wildness” , we urge to make use of it in teaching settings. This wildness allows to step out of conventional modes of critique, and into modes of action. We conclude with a protocol as a practical guide for everyone who wants to introduce data sprints in their teaching.