Numbers need words: How narrative takes your analysis from irrelevant to irreplaceable
?
3 likes?200 views
The document discusses the importance of providing context and narrative along with data analysis in order to create value for clients. It argues that data analysis alone is not sufficient and that stories must be told to explain the meaning, value and implications of the data as well as recommend next steps. The document provides tips for effective communication such as answering questions in order, understanding the client's perspective and constraints, and presenting findings in a clear and visually appealing way.
Numbers need words: How narrative takes your analysis from irrelevant to irreplaceable
- 2. By Barton Poulson Founder, datalab.cc Intermountain Big Data Conference Salt Lake City, UT 21 November 2015
- 3. Or, how narrative takes your analysis from irrelevant to irreplaceable.
- 6. 调产耻迟…皑
- 8. Data mojo × story = value.
- 10. {mojo} × 0 = 0
- 13. The meaning.
- 14. The value.
- 15. The plan.
- 16. The problems.
- 17. Egocentrism.
- 18. Jean Piaget.
- 20. False consensus.
- 22. Even in 1640.
- 23. Anchoring.
- 24. Broncos FTW!
- 25. Seahawks 43-8.
- 26. Probability.
- 27. 叠补测别蝉’罢丑别辞谤别尘.
- 28. Client != you.
- 29. Can’t crawl in.
- 30. What to do?
- 31. 1.Answer the questions. 2.Know the context. 3.Give the next steps. 4.Make it look good.
- 33. -State the question. -Give your answer. -Qualify as needed. -Go in order.
- 37. – Ludwig Mies van der Rohe, after Robert Browning Less is more.
- 39. – Albert Einstein If you can’t explain it simply, you don’t understand it well enough.
- 44. – Yogi Berra We’re lost, but we’re making good time.
- 45. -Unstated constraints? -Contradictory goals? -Client’s ability? -Social e?ects?
- 46. You have correlation. They want causation.
- 49. – William James My thinking is first and last and always for the sake of my doing.
- 50. -Recommend actions. -Goal oriented. -Be specific.
- 52. -Baby steps. -Doable by client. -Build on each step.
- 58. E ^ > <
- 60. Sorry, Mitt.
- 61. Make it flow. Make it make sense.
- 63. Well-presented good ideas are better than well-presented bad ideas.
- 64. Poorly-presented good ideas are indistiguishable from poorly-presented bad ideas.
- 65. Be creative. Be original. Even in data science. Especially then.
- 67. Myst.
- 68. Portal.
- 70. Asteroids.
- 71. Dwarf Fortress.
- 72. Canabalt.
- 73. Minecraft.
- 74. A tiny example.
- 78. Looks like bias. But wait… break it down.
- 80. But data isn’t enough. Look at context.
- 81. Why do programs vary in class size?
- 83. Why do men and women apply to di?erent programs?
- 85. -Identify departments. -Check admissions criteria. -Check prior education. -Check funding. -Adjust or raise funds?
- 86. Fire it up!