Strategies for the Analysis of Large Social Media Corpora: Sampling and Keyword Extraction Methods

Area of Science:

• Social Sciences
• Computational Linguistics
• Public Health

Background:

• Social media platforms like Twitter are crucial for information exchange during pandemics.
• Large-scale corpora from social media present data management challenges for researchers.

Purpose of the Study:

• To provide methodological and practical guidance for managing large social media corpora.
• To compare the efficiency and efficacy of different data handling techniques for social media data.

Main Methods:

• Comparison of different sample sizes and sampling methods for corpus management.
• Evaluation of traditional corpus linguistics keyword extraction versus graph-based Natural Language Processing techniques.

Main Results:

• Assessing the feasibility of achieving similar analytical results with varying sample sizes.
• Determining the effectiveness of different keyword extraction methods for representing corpus topics.

Conclusions:

• Effective strategies exist for quantitative and qualitative analysis of large social media datasets.
• The study offers practical solutions for researchers dealing with massive social media corpora.