Generate Insights, Not Excuses: How your data is slowing you down
summary
Argues that purposefully generating fresh, relevant data through experiments beats months of cleaning legacy logs, showing a workflow from expert-driven bayesian network prototype to fast deployment on the mindsmiths platform in under a month; the PCOS prototype collects clean signals and eventually becomes better than the world's leading expert at diagnosis. Recommends parallel A/B testing mass experimentation, notes usable models often emerge from just a few days of data, and cites ~90% project failure when clinging to messy historical data.
AI-written summary · about
chapters (6)
- 0:00Start
- 2:07Event Sourcing
- 9:21Data Analysis Is Hard
- 10:19The Lack of Relevant and Clean Data
- 21:18Making Concrete Data-Driven Intelligent Decisions
- 23:31Key Takeaways
related talks
47mData Analytics: Realizing the Full Potential
39mDoing Data: The Critical Process of Data Preparation (...)
23mBig data and big insights
35mBuilding Data-Intensive Teams
18mData quality: the key to long term happiness
41mProficiency + Creativity = Happy Data Practitioners
27mNavigating Data's Next Great Shift
24mValue-Driven Analytics Development