This directory helps to infer tweet's location using some inference techniques. (Still in development)

Requirement

• Node
• MongoDB
• Tweets and users collections in MongoDB (same as tweets and users object defined by Twitter)

How to run

• Follow the instructions and download Stanford NER at https://nlp.stanford.edu/software/CRF-NER.shtml
• Copy the stanford-ner-server.sh to Stanford NER's directory just downloaded
• Start the stanford-ner server inside Stanford NER's directory (you can choose one of three classifiers in the stanford-ner-server.sh before run the commands):

sh ./ner-server.sh

• Install dependencies (For the first time only):

npm i

• Start mongod server:

mongod

• Start infer your user location:

npm start user

• Start infer your tweet location:

npm start tweet

Credits:

• Stanford NER project: https://nlp.stanford.edu/software/CRF-NER.shtml
• Node client for Stanford NER: https://github.com/niksrc/ner
• Node google geocoder: https://github.com/nchaulet/node-geocoder