This paper introduces near-real time data on Russia’s full-scale invasion of Ukraine, and uses these data to investigate the short-term impact of occupation and violence on local economic activity. The data project – VIINA (Violent Incident Information from News Articles) – parses news reports from Ukrainian and Russian media, georeferences them, and classifies them into standard event categories (e.g. artillery shelling) through machine learning. As we show, VIINA is more geographically comprehensive and more thoroughly documented than other open-source event databases on Ukraine, and is the only such effort to track territorial control. We illustrate applications for research on political economy, by utilizing remote sensing data on luminosity and vegetation as proxies for urban economic activity and agricultural land use. We find that economic activity declined most in urban areas that neither side fully controlled, and in places most exposed to artillery shelling. Areas under Russian occupation, however, were more insulated from these negative shocks. Although not causal, these findings are in line with the view that war brings not only economic devastation, but also vast geographic inequalities that are almost immediately observable from space.

本文介绍了俄罗斯全面入侵乌克兰的近实时数据，并利用这些数据调查占领和暴力对当地经济活动的短期影响。该数据项目 – VIINA（新闻文章中的暴力事件信息） – 解析乌克兰和俄罗斯媒体的新闻报道，对它们进行地理参考，并通过机器学习将它们分类为标准事件类别（例如炮击）。正如我们所展示的，VIINA 比乌克兰的其他开源事件数据库在地理上更全面、记录更全面，并且是唯一追踪领土控制的此类数据库。我们通过利用光度和植被遥感数据作为城市经济活动和农业土地利用的代理来说明政治经济学研究的应用。我们发现，在双方完全控制的城市地区以及最容易遭受炮击的地区，经济活动下降幅度最大。然而，俄罗斯占领下的地区则更不受这些负面冲击的影响。尽管不是因果关系，但这些发现符合这样的观点：战争不仅带来经济破坏，而且还带来几乎可以从太空立即观察到的巨大地理不平等。