“…Computer scientists see this as a problem of mapping across database schemas, but in this case the target schema depends upon the content of the data, which cannot be specified in advance (Hernández et al, 2008). Several ways of formalizing and automating data-metadata translation have been proposed (Beine, Hames, Weber, & Cleve, 2014;Britell, Delcambre, & Atzeni, 2016;Wyss & Robertson, 2005;Xue, Shen, Nie, Kou, & Yu, 2013). The pivot-unpivot transformation is common in statistical analysis, where it is called "long to wide," and we link the choice between long versus wide data formats to data cultures found in different scientific domains.…”