Geolocation for Twitter.

A Python version is available here: https://github.com/mdredze/carmen-python

Carmen is a library for geolocating tweets. Given a tweet, Carmen will return Location objects that represent a physical location. Carmen uses both coordinates and other information in a tweet to make geolocation decisions. It's not perfect, but this greatly increases the number of geolocated tweets over what Twitter provides.

To compile: ant build

To create a new jar file (one already exists in the dist directory) ant jar

To clean: ant clean

The properties for controlling Carmen's behavior are location in: src/resources/carmen.properties

These parameters are explained in the javadocs for carmen.LocationResolver.

The paths in the properties file are currently relative paths. They should be updated to absolute paths as needed.

To run a demo: ant run-demo -Dargs='--input_file input.json'

You can provide a second argument to output the tweets with geolocation information to a file.

ant run-demo -Dargs='--input_file input.json --output_file output.json'

To run the experiments described in the Carmen paper (below). ant run-stats-demo -Dargs='--input_file input.json --output_file output.json'

input.json and output.json are both json files. input.json should contain tweets in json format, one per line. Twitter data is not distributed with Carmen. These files will be treated a gzip files if the have suffix ".gz"

To reference this package in publications:

@inproceedings{Dredze:2013a, Author = {Mark Dredze and Michael J Paul and Shane Bergsma and Hieu Tran}, Booktitle = {AAAI Workshop on Expanding the Boundaries of Health Informatics Using AI (HIAI)}, Title = {Carmen: A Twitter Geolocation System with Applications to Public Health}, Year = {2013}, }

README.md : this file LICENSE : 2-clause BSD license lib/ : dependencies src/ : source directory