UK charities that work internationally: an open dataset with validated SDG classification and a three-way overseas engagement flag
Jack Chen
No tky7s_v1, SocArXiv from Center for Open Science
Abstract:
One in ten charities registered in England and Wales says it works outside the UK. The register records where they work, but not what they work on or how: a water charity in Tanzania looks the same as a school in Nepal, and a charity running its own projects abroad looks the same as one that only raises money at home. I built an open dataset that fills both gaps for all 19,688 active internationally operating charities on the register, geocoded to their head offices. Each charity gets a primary UN Sustainable Development Goal, up to two secondary goals, and a three-way flag for how it engages overseas: operating directly abroad, funding partners abroad, or fundraising in the UK only. A large language model assigned every field from the charity's own register text. To check the tags, I hand-labelled a blind, stratified sample of 150 charities under rules fixed before labelling began. The model's primary goal matched my label 77.3% of the time (95% CI 70.0 to 83.3%), against a 56% benchmark from the nearest prior machine-learning work on the register. The engagement flag came out at 65.3%, and its errors lean one way, towards over-reporting overseas activity, so its most reliable call is "UK fundraising only". The dataset, code, validation labels, and an interactive map are public.
Date: 2026-07-21
References: Add references at CitEc
Citations:
Downloads: (external link)
https://osf.io/download/6a5b85ef0c5e23ba52e9fe69/
Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:osf:socarx:tky7s_v1
DOI: 10.31219/osf.io/tky7s_v1
Access Statistics for this paper
More papers in SocArXiv from Center for Open Science
Bibliographic data for series maintained by OSF ().