I'm a statistician at UNHCR, in the Secretariat of the Expert Group on Refugee, IDP and Statelessness Statistics (EGRISS), based in Copenhagen. I work with national statistical offices so that refugees, internally displaced and stateless people show up in official statistics, using the IRRS, IRIS and IROSS recommendations.
Most of the code here supports that work: the Global Annual Inclusion Survey (GAIN), SDG indicators for displaced people, and R pipelines that use a local LLM to find and check official statistics.
| GAIN SDG map | 110 SDG indicator values in 27 countries for forcibly displaced people, set against host populations. Code |
| GAIN evidence pipeline | R and a local LLM (Ollama, RAG) scanned 172 national statistics office websites and found 581 unreported examples in 108 countries. A validation layer rejects any figure the model can't tie to its source document. |
| Counted, and counting | Five years of GAIN implementation examples, 2021 to 2025, as one story page. Code |
| Identification questions builder | Prototype that assembles survey identification questions for a given country. Code |
| GAIN 2026 pilot desk | Tester desk for piloting the GAIN 2026 questionnaire. Code |
| GAIN Tables 2025 | Published tables, maps and analysis from the 2025 survey. |
R, Python, SQL, Power BI, R Shiny, Stata, Ollama, Git, KoboToolbox, SurveyCTO, CSPro, SDMX