Infrastructure · 14 projects · 2018 —
Not an end in itself, but a response to a concrete problem. After years of fieldwork across West Africa, I had accumulated thousands of documents that exceeded what traditional methods could process. These projects turn that accumulation into open research infrastructure — digital archives, AI-assisted pipelines, interactive visualisations, and conversational interfaces — so that African historical sources are not only preserved but made usable.
2023 — · Flagship
An open-access archive of over 14,700 items on Islam and Muslims in Benin, Burkina Faso, Côte d'Ivoire, Niger, Nigeria, and Togo.
14,700+ Items 6 Countries 4,600+ Index entries
2026 · New
Interactive research atlas for the Africa Multiple Cluster of Excellence — browse, filter, and compare projects, people, institutions, and research items across partner centres in Africa, Brazil, and Germany.
4,000 Research items 90+ Projects 42 Countries
A Model Context Protocol server and companion agent skill that give AI assistants structured, read-only access to the Islam West Africa Collection — overcoming the limitations of keyword search through structured, AI-mediated access to the corpus.
Interactive visualization of sentiment patterns in 12,280+ West African press articles on Islam, with multi-model comparison (ChatGPT, Gemini, Mistral) and third-party arbitration.
A reusable Docker template for deploying production-ready Omeka S digital archive installations with automatic setup, module management, and SSL support.
An interactive map tracing the public artworks of the Egyptian sculptor Hassan Heshmat across ten countries, developed for the Leibniz-Zentrum Moderner Orient (ZMO). Built with SvelteKit and MapLibre GL JS.
Python workflows for processing the Islam West Africa Collection using Large Language Models—automating OCR, NER, transcription, and summarization at scale.
The website for a VolkswagenStiftung scoping workshop held in Hanover in February 2026, which documents its programme and 26 participants and carries a filterable bibliography of over 200 works, a concept map of the literature, six filmed interviews, and the forthcoming position paper.
Google Colab notebooks that make AI accessible for researchers: transcribe audio/video, extract text from documents (OCR/HTR), and generate summaries—no installation required.
Two static web apps built from data scraped monthly off the ZMO website: a word cloud of what each research unit writes about, and a dashboard over the institute's full publication register — 1,962 publications, 656 authors, 1994-2026 — in which every chart is also a filter.
