Files
Download Full Text (4.5 MB)
SJSU Faculty Member
Vishnu Pendyala
Description
In an era where data never sleeps, streaming algorithms offer a powerful toolkit for extracting meaningful insights from high-velocity data flows. This talk explores some foundational techniques that enable efficient, real-time analytics with minimal memory requirements. The algorithms covered include a clever bit-based strategy for approximating the count of 1s in a sliding window, ideal for binary streams where space efficiency is paramount. Another algorithm helps estimate statistical moments (mean, variance, skewness) using compact sketches, enabling a deeper understanding of stream distributions without storing the entire dataset. One other algorithm identifies trending items with exponential decay, giving more weight to recent data, a crucial method for dynamic environments like social media or sensor networks. Techniques like these form the backbone of intelligent stream processing. Through intuitive examples and practical applications, this session will demystify how these algorithms work, why they matter, and how they can be used to monitor, summarize, and react to data in motion.
Publication Date
Fall 10-10-2025
Document Type
Presentation
City
San Jose
Keywords
Streaming algorithms, Real-time analytics, High-velocity data, Low-memory computation, Sliding window, Bit-based counting, Binary data streams, Compact sketches, Statistical moments from streams, Mean, variance, skewness, Exponential decay, Trending item detection, Data summarization, Stream distributions, Dynamic data environments, Social media analytics, Intelligent stream processing, Data in motion, Approximate computing
Disciplines
Computer Engineering
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Pendyala, Vishnu S., "Sensing the Pulse of a Data Stream in Real Time" (2025). Open Educational Resources. 17.
https://scholarworks.sjsu.edu/oer/17
