Python tools created and used as Wikipedian in Residence for Projekt Fredrika (projektfredrika.fi) to help edit over 10000 articles in Swedish Wikipedia, over 50000 Wikidata objects and thousands of Wikimedia Commons images. These scripts were created effectively thanks to the help of Generative AI.
- stats_wikipedia_multi_lang.py - used to compile statistics for Wikipedia article views over multiple languages, taking as input a Wikidata Q-code and fetching data from Wikipedia APIs
In projekt-fredrika's repo
- finsvetekniker.ipynb early AI-test for final project described at projektfredrika.fi/ai-sauna - Script that applies Generative AI to improve Wikipedia by analyzing sources (books), finding relevant Wikipedia articles to improve (with NER), and making improvement suggestions - using NER and LLM (OpenAI & Anthropic API)
- kotus-fo - Script to parse wordlist (XML) of Finland Swedish dialects published by Institute for the Languages of Finland (Kotus), and to transform it to be uploadable to Wikidata where possible (according to rules developed together with Kotus-experts)
- commons-statistik.py - used to compile view statistics for Wikimedia Commons images, fetching data from Commons' API