Overview
Explore this conference talk where Sourav Saha introduces FireDucks, a high-performance multithreaded DataFrame library with JIT compilation capabilities. Learn about the limitations of Pandas for modern data processing needs and discover how FireDucks addresses these challenges. The presentation covers best practices for efficient data processing, demonstrates FireDucks' implementation and usage through practical examples, and showcases its optimization features including multithreading and JIT compilation. Compare performance benchmarks between FireDucks and other DataFrame libraries, and gain insights into how this tool can significantly improve data processing workflows. The 42-minute talk provides a comprehensive overview from motivation to practical application, with resources for further exploration.
Syllabus
00:00 Introduction and Overview
00:38 Speaker Background and Experience
01:13 Motivation Behind FireDucks
02:34 Challenges with Pandas
04:37 Best Practices for Data Processing
14:35 Introduction to FireDucks
20:33 FireDucks Demo and Usage
26:24 Optimization Features of FireDucks
36:39 Benchmarking FireDucks
40:15 Conclusion and Resources
Taught by
Conf42