Tinkering away on my opensource Java NLP library to parse text and find geographical information from it.

https://github.com/tomaytotomato/location4j

I think I am going to re-write the logic to calculate a score on all matches it makes from a given piece of text.

e.g.

"us ca" ---> is this "USA California" or "USA and Canada (CA ISO2 code)"?

"san jose usa" ---> is this "San Jose California, USA" or another San Jose in America