mordecai

Why we included this project

Mordecai handles the full geoparsing chain in a single library call: it takes English text, finds the place names, and resolves each one to a specific location with coordinates and structured GeoNames data. Feed it "I traveled from Oxford to Ottawa" and it returns two correctly disambiguated places instead of one ambiguous match. It builds on spaCy for named entity recognition and an Elasticsearch-backed GeoNames gazetteer for resolution, so it fits into existing NLP and data workflows rather than asking you to assemble the pieces yourself. That makes it a practical fit for event data, news analysis, or any corpus where locations matter and the structured output is easy to consume downstream. One caveat: it needs a running Elasticsearch instance with the GeoNames index loaded, so plan for that dependency before adopting it.