Abstract
The Measurable Truth pilot dataset presents systematic quantitative coding of 13 visual investigations (typically presented as online linear videos or scrolling narratives concerning humanitarian crises) produced by Airwars, an open-source intelligence (OSINT) organisation specialising in civilian harm monitoring, between 2021 and 2025. Each investigation is coded across 46 variables encompassing methodological technique composition (expressed as four proportional ratios: Observational, Computational, Hybrid, and Synthetic), presentational format, circulation context, methodological reflexivity, and collaborative structure. The data was collected through structured manual close-reading of publicly accessible investigations, using a controlled vocabulary of 27 technique types. This is an initial pilot corpus designed for expansion: it forms part of ‘Digital Verite’, a broader research project examining the changing conditions of knowledge production and truth value in computational visual documentation across OSINT, journalism, and human rights practice. The pilot dataset supports cross-institutional technique comparison, analysis of the relationship between investigative format and epistemic transparency, and longitudinal tracking of methodological change in the visual investigation of conflict and harm. It is deposited as a structured spreadsheet with JSON and CSV exports, and is intended for aggregation with equivalent datasets from other institutions.
© 2026 Isaac Parkinson, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.