In the last decade, big textual datasets in the humanities have become increasingly more available in the form of raw data. The challenges these datasets raise are twofold. On the one hand, most humanities scholars are not equipped with skills in text and data mining. This remains a barrier to study big data in the humanities. On the other hand, the traditional genre of digital edition is not suitable for publishing and unlocking big data; similarly to printed editions, digital editions often attempt to create highly curated, almost perfect, surrogates of texts with critical accuracy. However, in the context of big data, traditional critical accuracy is not attainable; it is impossible for an editorial team to apply this principle when working with a corpus of tens of millions of words. The principle of critical examination of texts as defined by previous scholarship is equally unattainable with big data. In short, many of the editorial principles and techniques used to produce analogue and digital editions can hardly be applied when creating an edition featuring truly big data. Hence, in this chapter, I argue that to make big data available and explorable for the scholarly community, we need a new genre: the scholarly data edition. Throughout the chapter I elaborate the concept of scholarly data edition by outlining the editorial responsibilities and standards that it involves.