Managing Circularity in Data Product Development
summary
Argues data work splits into linear vs circular projects and that circular work often has unknown outcomes — sometimes after years of feature engineering a crowdfunding risk model still performs no better than the baseline average. Recommends concrete practices: invest in requirements, run time‑boxed spikes, use design sprints, and aim for the smallest meaningful iteration. Offers guidance on setting explicit timeboxes (two/ four/ six weeks) and stopping rules so projects don’t run forever.
AI-written summary · about
chapters (8)
- 0:00Intro
- 1:07Data work is hard
- 7:30Linear Projects
- 8:49Circular Projects
- 16:40Classify Projects
- 17:18Set Expectations
- 21:27Team Process - Goals
- 28:40Managing Circularity
related talks
40mData-Driven Agility
26mData For The 99%: Unlocking Data Curiosity Across Your Organisation
41mBreaking the 80/20 Dilemma in Data Science Projects
40mPanel Discussion: What Users Really Want
48mBecoming a Data Head, a reflection on the trade offs and decision
37mEngineering data quality
27mVisual Design - Do not let Data be your cul-de-sac!
32mWhy Most Data Projects Fail and How to Avoid It