Data science done wrong: Dissociation and disempowerment
3 likes183 views
Barton Poulson discusses the potential pitfalls of data science, emphasizing the risks of dissociation and disempowerment that arise when data is regarded as the sole truth. He critiques the tendency to rely on data without context and warns against oversimplified conclusions that can lead to misguided policies and decisions. The document encourages a balanced approach to data interpretation, advocating for healthy skepticism and the consideration of human experience in data analysis.
Data science done wrong: Dissociation and disempowerment
- 1. Data science done wrong. Dissociation & Disempowerment.
- 2. By Barton Poulson Founder, datalab.cc Intermountain Data Conference Salt Lake City, UT 19 November 2016
- 4. I like data science. I like what it can do. I like it best when it’s done thoughtfully.
- 5. The problem starts. We’re smart, we have computers, we have algorithms.
- 6. Famous last words. What could possibly go wrong?
- 7. Hello, Dave.
- 9. AI defense.
- 10. AI offense.
- 11. So much for data.
- 12. WC Fields “Hubris is back in town.”
- 13. Dissociation?
- 14. Dissociation. The experience of having one’s attention & emotions detached from the environment.
- 15. “Gaslight”
- 16. Data vs reality.
- 17. A horse story.
- 19. In the year of our Lord 1432, there arose a grievous quarrel among the brethren over the number of teeth in the mouth of a horse.
- 20. For thirteen days, the disputation raged without ceasing. All the ancient books and chronicles were fetched out, and wonderful and ponderous erudition, such as was never before heard in this region, was made manifest.
- 21. At the beginning of the fourteenth day, a youthful friar of goodly bearing asked his learned superiors for permission to add a word, and straightway, to the wonderment of the disputants, whose deep wisdom he sore vexed, he beseeched them to unbend in a manner coarse and unheard-of, and to look in the open mouth of a horse and find answer to their questionings.
- 22. At this, their dignity being grievously hurt, they waxed exceedingly wroth; and joining in a mighty uproar, they flew upon him and smote his hip and thigh, and cast him out forthwith. For, said they, surely Satan hath tempted this bold neophyte to declare unholy and unheard-of ways of finding truth contrary to all the teachings of the fathers.
- 23. After many days of grievous strife, the dove of peace sat on the assembly, and they as one man, declaring the problem to be an everlasting mystery because of the grievous dearth of historical and theological evidence thereof, so ordered the same writ down. - Francis Bacon (1561-1626)
- 24. The risk. The belief that if it isn’t data, it isn’t real.
- 25. Disempowerment?
- 26. Disempowerment. To deprive of power, authority, or influence; make weak, ineffectual, or unimportant.
- 27. Data science results are not written in stone.
- 28. Not like this.
- 29. And you don’t want to hand over your life to a kiosk.
- 30. Or this either.
- 33. An oversimplification. • IQ is measureable. • IQ is important. • IQ is heritable. • IQ is immutable.
- 34. But then it suggests… • IQ is lower in poor people, immigrants, and minorities. • IQ should drive policy.
- 36. Remember that slam dunk election of 2016?
- 37. Projections
- 39. “Data died tonight.” {GOP strategist Mike Murphy}
- 40. How many people stayed home because they thought it was a done deal?
- 41. Subject & object. One thing when it’s you… But different when it’s other people.
- 42. To do or to be done to.
- 43. What to do?
- 45. Sturgeon’s Law. Ninety percent of everything is crap.
- 46. Standards of proof. Extraordinary claims require extraordinary evidence.
- 49. Bayes’Theorem.
- 50. Use your priors. New data should be considered in context of priors, including personal experience.
- 51. One hard data point.
- 53. You know the Data Science Venn Diagram? {Thanks, Drew Conway}
- 55. Or the proposed fourth circle? {Thanks, Michael Malak}