"Data Science & Agile: Strategies for Working Together"
summary
Data science and Agile collide over timelines, access, and measurement; concrete fixes include using research spikes or a month‑long proof of concept instead of forcing two‑week feature commits, changing deliverables to regular status updates rather than “finished code,” and freezing planning to one or two iterations ahead. Real outcomes: an initial image classifier stalled by conservative commit but would have progressed faster with result‑focused updates, and a data‑entry automation reached >90% accuracy but only proved valuable once compared to human error rates and time savings.
AI-written summary · about
chapters (13)
- 0:00Welcome
- 1:45Data science challenges
- 7:22Agile principles for data science
- 9:07Agile vs data science complaints
- 11:17Case study: Research timelines
- 13:27Planning limits in data science
- 14:30Communication gaps
- 16:18Case study: Data access alignment
- 18:03Agile overhead issues
- 20:07Metrics disconnect
- 21:31Case study: Human vs machine accuracy
- 22:57Collaboration strategies
- 24:00Q&A
related talks
40mAgile Software Development: Practice and Analytics
58mDelivering Agile BI Solutions
57mContinuous Discovery: The Power of Pure Agile
39mScrum at Scale
34mHow to screw Agile and digital transformation
01h 33mPanel - Predictability and Productivity...Does Agile Help?
26mGet agile and don't die trying
11mLeading Agile Teams to Success: Lessons from the Trenches