Modern Analytics: One Size Does Not Fit All

on YouTube ↗26m

" summary

Argues that one size does not fit all for analytics by pinpointing memory, storage and data models as the unseen forces reshaping platforms; shows how columnar formats (Parquet/ORC) and tools like Presto/Hive fail for highly normalized ERP data but how in-memory join indexes and an agile data pipeline let companies go from hours to seconds. Examples: close books 5 days earlier, 40% cut in food production costs across 28,000 stores, 35 billion transactions handled, 2,000 SKUs repriced in 16 hours.

AI-written summary · about

" chapters (10)

1. 0:00Start
2. 2:42First Custom-Built Solutions
3. 3:23Parthian Battery
4. 5:31Recap
5. 6:24Ubiquitous Cloud Computing
6. 10:07Performance Benefits
7. 11:07Data Models
8. 20:14Headless Bi
9. 22:38Mantra of Being Good Enough
10. 25:16Parting Thoughts

" related talks

• 10mImpacting Business as it HappensAnil Gadre · Strata + Hadoop World San Jose 2015
• 19mThe Challenges of Product AnalyticsTamas Spancik · DSC Europe 2022
• 26mColumnar Storage: Redefining Data Management for the Modern EraZoe Steinkamp · Voxxed Days Thessaloniki 2024
• 01h 01mModern Data Warehousing StrategyMarco Russo · PASS Business Analytics Conference 2013
• 01h 00mBig Data, Bad AnalogiesMark Madsen · GOTO Copenhagen 2014
• 47mData Analytics: Realizing the Full PotentialMike Stringer · PASS Business Analytics Conference 2013
• 21mPutting big data self-service analytics in the hands of business peoplePeter Kacandes · Big Data & AI Toronto 2020
• 43mAnalytics capabilities architectural patterns: A journey for data-driven insightsJohn Mousa · Voxxed Days Brussels 2023