Derivation & Presentation: How to Effectively Tell A Data Science Story
summary
Claim: data science succeeds only when it prescribes actionable change, not by running models for their own sake. Time in practice splits into prep/quality, analytics (about one third of effort), and presentation; stakeholder attention lasts 5-10 minutes. Work uses IPython notebooks, scikit-learn, matplotlib/seaborn and D3 for visuals; validate hypotheses with a quick notebook and a stakeholder check to avoid wasted work. Examples include productized models like Spotify recommendations and ad hoc KPIs that drive roadmap shifts when conclusions are stated as main points.
AI-written summary · about
chapters (11)
- 0:00Storytelling with data science
- 3:32Data science and business role
- 7:08Data scientists' real-world time
- 14:14Derivation: Finding your mountain
- 19:53Validating insights before deep work
- 21:41Prescribing actionable change
- 24:08Communicating insights effectively
- 28:00Tuning tone and abstraction
- 33:20Pictures and metaphors for presentations
- 35:49Preparing for presentation curveballs
- 36:52Guiding audience through presentation
related talks
36mKeynote: Data, Decision Making, and Being Human
55mThe Art Of Business Storytelling
34mKeynote: Lighthearted Thoughts on Conversing About Data
21mSpeaking the Same Language
31mStorytelling for Engineers
22mCultivating creativity in data work
07mEmotional Healthcare and Data Storytelling
01h 00mHow to Pitch Your Ideas so People Listen