A general framework for event detection from social media

Khatereh Polous, André Freitag, Jukka Krisp, Liqiu Meng, Smita Singh

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The availability of accurate and/or up-to-date mass data can stimulate the development of innovative approaches for the assessment of spatio-temporal processes. However, extracting meaningful information from these collections of user-generated data is a challenge. Event detection is an interesting concept in the era of Web 2.0 and ubiquitous Internet. Various existing event-detection algorithms share a very simple, yet powerful architecture model; pipes-&-filters. Using this model, the authors in this study developed a generic and extensible programming framework to find meaningful patterns out of heterogeneous and unstructured online data streams. The framework supports researchers with adapters to different social media platforms, optional preprocessing steps. Its graphical user interface supports users with an interactive graphical environment for setting up parameters and evaluating the results through maps, 3D visualization, and various charts. The framework has been successfully tested on Flicker and Instagram platforms for different time periods and locations to detect latent events.

Original languageEnglish
Title of host publicationAdvances in Spatial Data Handling and Analysis - Select Papers from the 16th IGU Spatial Data Handling Symposium
EditorsYee Leung, Francis Harvey
Publisherspringer berlin
Pages85-105
Number of pages21
ISBN (Print)9783319199498
DOIs
StatePublished - 2015
Event16th International Symposium on Spatial Data Handling, SDH 2014 - Toronto, Canada
Duration: 6 Oct 20148 Oct 2014

Publication series

NameAdvances in Geographic Information Science
Volume19
ISSN (Print)1867-2434
ISSN (Electronic)1867-2442

Conference

Conference16th International Symposium on Spatial Data Handling, SDH 2014
Country/TerritoryCanada
CityToronto
Period6/10/148/10/14

Keywords

  • Datastream mining
  • Event detection
  • Flicker
  • Instagram
  • Knowledge discovery
  • Social media

Fingerprint

Dive into the research topics of 'A general framework for event detection from social media'. Together they form a unique fingerprint.

Cite this