Why OpenData Doesn't Exempt You from Some Statistical Basics
summary
Guillaume Rozier and Sacha Guilhaumou examine how France’s OpenData (used in projects like CovidTracker) provides abundant up-to-date pandemic numbers but does not eliminate the need for statistical literacy. They show specific examples of statistical paradoxes and biases that can lead to misleading or false conclusions when interpreting COVID data, and explain precautions needed when analyzing such datasets.
AI-written summary · about
chapters (12)
- 0:00Presentation of the topic
- 3:20Introduction of G.Rozier
- 7:11Classic visual illusions
- 10:36Classic statistical errors
- 12:12Intersection and conditional probability
- 13:23Bayes' theorem
- 15:30Confirmation bias
- 16:35Simpson's paradox
- 19:34Age-related bias
- 21:05Small sample size bias
- 23:00Summary
- 23:48Questions
related talks
38mRole of Statistics in the era of Big Data
27mDemocratizing Data: A deep dive into Eurostat Open Database
37mHow to Start Critical Thinking in Data Science
01h 19mOpportunities and risks of big data in disease surveillance.
43mThe Subtle Science of Misleading with Statistics
01h 10mThe Social Survey Statistician's Perspective
27mThe Subtle Art of Lying with Statistics
48mCommon Data Science Mistakes