The dataset underlying the second-wave paper (“Understanding the effectiveness of government interventions against the resurgence of COVID-19 in Europe”): a structured record of non-pharmaceutical interventions against SARS-CoV-2 across European regions.
See also: npi-second-wave.md.
Converted from the open-access Nature Scientific Data descriptor (Europe PMC, PMC8975844). Markup removed; the exhaustive data-dictionary and reference list are condensed. This is a Data Descriptor, so it documents the dataset rather than reporting new findings.
Abstract. During the second half of 2020, many European governments responded to the resurging transmission of SARS-CoV-2 with wide-ranging non-pharmaceutical interventions (NPIs). These efforts were often highly targeted at the regional level and included fine-grained NPIs. This paper describes a new dataset designed for the accurate recording of NPIs in Europe’s second wave to allow precise modelling of NPI effectiveness. The dataset includes interventions from 114 regions in 7 European countries during 1 August 2020 to 9 January 2021, with NPI definitions tailored to the second wave following an exploratory data collection. Each entry has been extensively validated by semi-independent double entry, comparison with existing datasets, and, when necessary, discussion with local epidemiologists.
First-wave NPIs were generally national and analysed with public datasets; the second wave saw local, fine-grained NPIs (small gathering-size bans, sector-specific business closures, curfews) that first-wave definitions were too coarse to capture. This dataset’s distinguishing features: precise NPI categorisation (greater differentiation in gathering size, gathering type, and business-closure types), regional granularity (for a period when NPIs were often subnational), and manual collection with semi-independent validation (higher quality than purely automated collection). It underpins Sharma et al. (the second-wave effectiveness paper) and has been reused to adjust for NPIs when analysing seasonality and mask use. Researchers should choose it when they need subnational data for this period, per-entry context (comments, quotes, sources), second-wave-tailored definitions, or fully validated data.
Design process. An open-ended exploratory collection first identified the main second-wave NPIs, suitable cross-country definitions, and the right administrative level per country — revealing that first-wave definitions were too coarse (e.g. the strictest gathering ban elsewhere is “≤10”, but many key second-wave NPIs were limits of 6/4/2 or a full ban). Design decisions were then finalised with team and external-expert input.
Coverage. 7 countries, 114 regions of analysis, 1 Aug 2020 – 9 Jan 2021, >5,500 entries. Administrative divisions: whole-country for Austria (9 states), Czechia (14 regions), Italy (21 regions), Netherlands (25 safety regions); stratified random samples of 15 for England (NUTS-3), Germany (districts, from the 4 largest states = 60% of population), and Switzerland (cantons). Regions of analysis = the largest division with uniform NPIs; regions with <2,000 cases excluded. Countries were included only where daily case/death data were public at the NPI implementation resolution (excluding e.g. Spain’s ~8,000 municipalities).
NPIs recorded. Public/private and indoor/outdoor gathering limits (exact person- and household-thresholds); business closures (night clubs, gastronomy, leisure/entertainment, retail + close-contact services, all non-essential — closures only, not safety measures); primary/secondary school closures (incl. holidays and most-teaching-online; for universities, vacation and student-dispersal periods); the OxCGRT mask-stringency scale (0–4); and curfews. Stay-at-home orders were excluded (too many second-wave exceptions, often only in legislation), as was shielding (no consistent cross-country definition).
Data collection & validation. Nine researchers, ~950 hours (185 on national timelines, 470 on regional data, 290 on validation). Each entry has dates, sources (government/legal/ media), quotes, and comments. Guiding heuristics: record an NPI when it affects “most or all” (>50%) of the relevant institutions/events/people; record from when behaviour is expected to change; record only mandatory NPIs (with a few documented exceptions for private gatherings confirmed via local epidemiologists). Validation: comparison against OxCGRT (and additional external sources for Italy/England); semi-independent double entry (a second researcher re-entered dates with access to sources/quotes but not the original dates; differences auto-flagged and resolved); cross-country consistency checks; and automated checks (dates in range, chronological order, no spurious gaps).
Data records. Published on figshare — a human-readable CSV per country, plus a machine-readable CSV (a row per day 1 Aug 2020 – 6 Jan 2021 with columns per NPI; ~19,000 rows across 144 regions) including daily cases/deaths. Recommended alternatives exist for researchers needing more countries, greater geographic/income variety, the full pandemic period, or economic measures (not included here).