NLP for GIS
summary
Uses a licensed Airbnb dataset of millions of listings and billion words to map cross‑border property interest and domestic flows, producing pixel maps that reveal capital to coast links and resort hotspots. Geocoding relies on an OpenStreetMap names gazetteer. Scales named‑entity search by reengineering search with an Aho‑Corasick trie and handles misspellings with fast fuzzy matching using n‑grams + TF‑IDF + ANN. Code and a Kaggle data snapshot are slated for release.
AI-written summary · about