See if talking to the doctor is right for you
summary
Clinicians hand over messy Excel exports and expect binary answers; data scientists want uncertainty and diagnostics. The session shows working R examples (Posit ecosystem mentioned), demonstrates why dichotomizing continuous variables ruins inference, points out a reported odds ratio CI spanning 7–17 as unhelpful, and explains bootstrap confidence intervals and QQ plots used to decide fixes. Practical advice: learn each other's IDEs, ask why a model or transform is chosen, and program analyses together.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:34Clinicians versus data scientists
- 2:58Doctors learning biostatistics
- 4:04Continuous variables and error
- 6:15Confidence intervals and uncertainty
- 8:22Partnership through nosiness
- 10:34Diagnostics and therapeutics
- 12:49Clinicians' view of data science
- 13:33Closing and thank you
related talks
44mPanel: Career Advice for Data Scientists
22mWhy is a good data scientist a package of professions?
44mHow to Teach an Old Firm New Tricks
34mKeynote: Lighthearted Thoughts on Conversing About Data
22mWhat Developers and Economists Can Learn from Each Other!
20mR! You Going?!
38mGetting ready for a career in Analytics and Data Science
18mData Science Careers via A/B Testing