Show HN: EDA Benchmark Leaderboard

Category: ai-ml

Tags: benchmark, synthetic-data, agents, data-analysis, evaluation

Score: 4.5/10 (Innovation: 5, Technical: 4, Documentation: 6, Utility: 3)

EDA Benchmark is a public leaderboard and dataset repository for evaluating AI agents on synthetic data-analysis tasks, emphasizing repeatability via a reliability-adjusted score. It offers 10 diverse example tasks (e.g., anomaly detection, data cleaning) and benchmark code, though its primary purpose appears to be marketing for deepsense.ai's custom dataset services.

Target audience: AI researchers, data scientists, ML engineers

Repository: https://deepsense.ai/blog/eda-benchmark-leaderboard-july-14-2026-update/ · Python · NOASSERTION · 18 stars

View on Hacker News