Altmetrics. New measures for the visibility of research results

The measures traditionally used to analyze and evaluate the use of scientific production are based on citation analysis using large corpora of document references. In recent years, the digital environment has opened up new possibilities for measuring scientific activity online. These new metrics cover both new scientific objects (datasets, software, videos, etc.) and new data sources, notably those of social networks (Mendeley, Facebook, Twitter, etc.), hence their original name of "altmetrics" as alternative metrics to those used until now.

It is important to understand that the notion of "alternatives" does not mean that they have the ambition to replace traditional metrics, but rather to provide new types of information on new forms of online activity (discussion, sharing, recommendation, etc.) around research productions taken in the broadest sense, i.e. no longer just journal or conference articles and books, but also all those that did not previously exist or were not taken into account by traditional metrics. What's more, beyond the informational dimension of attention measurement, the tools and services developed around these metrics can also prove to be genuine filtering and navigation tools in the mass of scientific resources for researchers, or strategic intelligence instruments for institutions and research funding bodies.

As the field of alternative metrics is still in its infancy, its development is largely the subject of varied and, in some cases, discontinuous experimentation. In this particularly evolving landscape, a number of players are already positioning themselves and consolidating around differentiated strategies, in terms of economics, services developed and target audiences. However, the potential for development remains particularly strong, and there are still many difficulties, both social and technical, to be resolved when it comes to collecting and reusing data. The data compiled and the sources from which they are derived are highly heterogeneous and unstabilized in their evolution, making comparisons both synchronic (of the different services and the measures they offer) and diachronic (of the sources and the data themselves) difficult, if not impossible. The use of social networks is also still far from widespread in scientific communities, and involves numerous biases linked to linguistic and disciplinary areas. What's more, the audience measured in this way is not limited to these scientific communities alone, which complicates the differentiation between scientific impact and social impact, and increases the risks of audience manipulation. Finally, it is important to distinguish between the measurement of impact and that of quality, which cannot be defined by audience criteria alone. The future of these new metrics therefore remains...