The increasing proliferation of social networking platforms has led to an unprecedented volume of dynamic and diverse data, posing significant challenges for traditional and contemporary big data processing technologies. Traditional systems, primarily designed for structured, static data, struggle to handle the multifaceted and unstructured nature of social media data. Contemporary solutions like Hadoop’s MapReduce, while capable, exhibit performance bottlenecks due to intensive I/O operations on disk storage. This paper explores the viability of Apache Spark as a robust alternative for social media data analytics, addressing the shortcomings of existing technologies. Spark’s in-memory processing capabilities and extensive libraries offer substantial performance improvements and flexibility, making it well-suited for real-time data processing and complex analytics. Through detailed use cases, including product enhancement via review analysis and marketing optimization through behavioral insights, the paper demonstrates Spark’s potential to transform social media data analytics. The study concludes with a discussion on future work, emphasizing the need for practical implementations to quantify Spark’s efficacy in real-world social media data scenarios.
Kaushal Thaker (Wed,) studied this question.