Making Data Relevant to Business. It's Harder Than You think!
summary
Prototype uses scikit-learn random forest and simple decision trees, achieves 85% accuracy in experiments but deploys as a batch double-check reducing misses 10%→8% (≈$60k/year). Implementation needs two weeks of analyst training, little engineering, no streaming infra, and runs safely behind human QA so risk and legal exposure stay low. Jupyter notebooks stay for development; a concise PowerPoint plus clear metrics drives adoption.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 7:59Data science project failures
- 12:51Key considerations for data projects
- 16:40Avery's data strategy case study
- 24:46Business-friendly presentations
- 30:22Soft skills for data pros
- 37:22Resources for business skills
- 45:23Q&A
related talks
36mDigital, data-driven financial advisory
39mKeynote Panel State of AI in the Enterprise
30mOperationalizing Machine Learning
13mEnterprise-Level Data Science Success- 13mEnterprise-Level Data Science Success
28mHow to manage data-related projects and not fail (too often)?
35mBuilding Data-Intensive Teams
22mBuild Smarter, Don't Just Build