Role of Statistics in the era of Big Data
summary
Argument that classical statistical principles remain essential in Big Data, with concrete historical examples: opinion polls fail due to non‑representative samples (social media and phone polls), flawed election and sports predictions from misuse of averages and biases, and clinical‑trial errors from ignoring patient heterogeneity. Claims need for explicit methods—sampling design, regression and correlation diagnostics, quality‑control checks—and a practical prescription to combine those with modern AIML pipelines and large compute to avoid large, preventable errors.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 2:14Big Data challenges
- 10:42Data-driven techniques
- 22:20Statistics in predictions
- 29:55Statistical ideas history
- 36:10Q&A summary
related talks
26mData Bias: The Danger Lies in the Details
07mNew Technologies with Big Data
37mThe Laws of Classical Physics Also Apply to Big Data!
55mKeynote - Big Data's Ethical Drought
01h 31mBig Data: The Computation/Statistics Interface
02mDBTB INT S Srinivasan r
46mWhat's the Science in Data Science?
38mDo's and don'ts of big data analytics