Data Science Project Patterns that Work
summary
Practical patterns for the five project stages: choose good projects, de‑risk data by talking to users and using schemas and profilers, do productive research with diagnostics and tools like Pandera pandas‑profiling Lux SHAP and LIME, deliver early with simple rule‑based E1s or CSVs, and iterate until users trust and adopt the system. Includes a concrete insurance story of walking the underwriting process, concrete anti‑patterns (hacking notebooks, handoff deployments) and concrete tactics to ship repeatable value.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 3:34Choosing good projects
- 7:03Understanding data meaning
- 12:14Diagnostic tools for research
- 17:50Iterative deployment for value
- 21:41Building trust in organizations
- 25:14Successful patterns summary
- 27:19Q&A
related talks
40mBuilding Successful Data Science Projects
43mAgile Strategies for High-Impact Data Science & Machine Learning Projects
39mDoing Data: The Critical Process of Data Preparation (...)
29mData Science Best Practices : From Proof of Concepts to Production
25mBuilding A Successful Data Science Team
28mTaking Data Science in industry from zero to production
32mWhy and How to Craft Bespoke Data Science
44mCareers in Data Science