CrowdSourced weather reports: An implementation of the µ model for spotting weather information in Twitter

Abstract

Twitter is a microblogging facility that allows people to post 140 character status updates about various topics. In times of special events (such as extreme weather, emergencies, sporting goals, etc), status updates on Twitter often give people a better view of the event than traditional news operations or weather services. This paper describes a project in monitoring Twitter for weather status updates for a specific city and being able to automatically determine the current weather by analysing those tweets.