This is the source code for the paper "GeoBurst: Real-Time Local Event Detection in Geo-Tagged Tweet Streams".
The usages of different modules are listed below:
-
demo: contain the main class file for running the GeoBurst method
-
utils: some general utility functions
-
geo: some utility functions for geo-tagged tweets
-
graph: calculate the inbound rwr score based on the keyword co-occurrence graph
-
clustream: construct the activity timeline using the clustream algorithm
-
hubseek: clustering geo-tagged tweets into geo-topic clusters
-
rank: ranking candidate events