The paper examines the methodological aspect of using data in historical research. In the context of dataism, discussions are presented about the features of formalizing information from historical sources in data models. To-day, in debates about the principles of using information technologies in historical research, one can encounter many contradictions related to the complexities of interaction between the approaches of computer, historical and social sciences. There are still oppositions between qualitative and quantitative approaches, despite the widespread use of “mixed methods” and Digital Humanities research approaches, as well as the differences between the goals of the information technology industry and academia. Data becomes the commonplace of such discussions because actual research relays on it. The paper examines the features of data in historical research, which make it possible to abstract from the original historical sources and collect systematic, formalized observations. Such data features as format, memory and passivity are considered. Attention is paid to such properties of research data as volume, velocity, varie-ty, veracity, variability, visualization, value (7Vs). It argues that historical data, in a semiotic sense, performs three functions: naming the properties of objects in the real world (nomination), connecting the named properties (predica-tion), and locating the named objects in space and time (location). From a methodological perspective, data in histor-ical research is viewed as a method of abstract observation. It involves systematically and strictly formalizing obser-vations from various historical sources. Despite the established disciplinary differences in digital approaches within history and digital humanities, data becomes a common basis that expands the researcher’s toolbox.