About

NOISE is a data structuring project developed alongside a dissertation funded by the Swiss National Science Foundation.

The project centers on the systematic extraction and consolidation of named entities—primarily persons and companies— from a wide range of historical sources, including directories, membership lists, trade indices, and archival records. Individual occurrences are parsed and normalized into entities, allowing actors to be traced across time, sources, and institutional contexts.

This approach enables longitudinal analysis of economic participation and stability: memberships, affiliations, and activities can be followed across decades, compared between datasets, and related to specific locations and sectors. Most entries are geocoded, and trade activities are categorized, allowing the data to be explored both spatially and economically.

While some sources contain explicitly relational information, NOISE does not impose a fixed network model. Instead, connections emerge through aggregation: patterns of overlap, persistence, exception, or absence become visible only when entities are viewed across the full corpus.

The name NOISE reflects this methodological stance in a second sense. Beyond the acronym, it points to the “noise” of complex economic realities—the minor actors, irregular patterns, and unexpected constellations that are often filtered out in traditional narratives focused on large firms or formal institutions. By structuring data at the level of individual actors and their occurrences, the project enables the study of these subtler dynamics within global trade networks, particularly under the conditions of crisis and war.