A Review on the Trends in Event Detection by Analyzing Social Media Platforms' Data

Area of Science:

• Computer Science
• Social Computing
• Data Science

Background:

• Social media facilitates collective action and information sharing.
• Misuse of social media for harmful events necessitates robust detection mechanisms.
• Event detection from social media data is crucial for societal security.

Purpose of the Study:

• To systematically review and analyze research on social media event detection.
• To identify current techniques, tools, and datasets in the field.
• To provide guidelines for future research directions.

Main Methods:

• Systematic literature review of 67 articles from the past decade.
• Analysis of event detection techniques, tools, technologies, and datasets.
• Evaluation of performance metrics and identified challenges.

Main Results:

• Reviewed papers focus on detecting natural disasters, traffic, sports, and real-time events.
• Key challenges include language compatibility (spelling, dialects) and multi-modal data processing (text, image, video, location).
• Existing event detection techniques lack robustness for heterogeneous data and languages.

Conclusions:

• Current event detection algorithms are not fully equipped to handle diverse languages and media formats.
• There is a need for advanced event detection techniques that are compatible with heterogeneous data and platforms.
• Developing 24/7 real-time detection systems integrating event and location data is essential for improved performance.