A world of data
summary
Argues that observational data at web scale enables fast learn-and-refine cycles: news pages move from editor curation to algorithmic selection using multi-armed bandits, backed by offline Hadoop models and low-latency NoSQL stores like PNUTS. Gives concrete wins — a school district lifts graduation rates from 55% to 78% — and cites IoT and internal systems managing exabytes with hundreds of petabytes/day I/O and 40–50k-node clusters; mentions Cosmos rebuilt for HDFS/YARN and GDPR operational needs.
AI-written summary · about
chapters (8)
- 0:00Introduction to data revolution
- 1:57Personalizing web content at Yahoo
- 6:13Next wave of data explosion
- 9:04Cloud democratizing data tech
- 10:05Cloud innovations for data apps
- 13:40Microsoft’s data platform strengths
- 14:44Cosmos: Microsoft’s big data backbone
- 17:37Privacy and compliance in data systems
related talks
41mOn Data: Sky above, sand below, peace within
37mTaming the Data Tsunami
32mWhy every developer should be a data scientist
36mWhen the Only Way to Scale up, is to Scale Down
44mPeople, Technology, Culture: How to make Data Science work
36mBig Data, Disruption And The 800 Pound Gorilla In The Corner
42mTales from the Field: The Future is Now
48mHyper Speed! When Big Data Blooms