The “science of big data” and data-driven journalism are frequently linked because they arise from the same effort to incorporate data into the practices of producing information. However, they stem from epistemological approaches that are difficult to reconcile with one another. While some find that big data, with its incommensurable datasets, has a monopoly on correlation without causation, data journalism appeals to standardized data to justify the truth of its statements. From this perspective, data must be reprocessed and tested in the field in order to become meaningful. Data does not speak for itself, as can be seen in the first great investigations that marked the early days of data journalism.