The Real World of Data Analytics

on YouTube ↗38m

" summary

Argues that projects fail on plumbing not models, citing an IBM Watson oncology prototype halted after $62 million / 5 years when MD Anderson switched EMRs and broke a rigid warehouse; proposes a flexible semantic approach using URI/RDF semantic data lake that runs in production at Montefiore for two years. Gives concrete algorithms: a random forest with 46 covariates to predict respiratory failure (estimated savings ~$4M/year) and NASA’s ISHM for early sensor-failure detection at Kyocera.

AI-written summary · about

" chapters (4)

1. 0:00Start
2. 3:45Typical Machine Learning Processes
3. 29:55Inductive System Health Monitoring Algorithm
4. 33:01Summary

" related talks

• 41mUsing Data Riches: A Tale of two projectsAjay Vinze · Data Science Conference 4.0
• 19mA world of dataRaghu Ramakrishnan · Collision 2017
• 32mThinking Like a Data ScientistEm Grasmeder · Lambda World Cádiz 2019
• 07mOpening WelcomeJulie Steele, Colin Hill · Strata Rx Conference 2013
• 51mFrom Data Confusion to Data IntelligenceElaine McVey, David Meza · posit::conf(2023)
• 32mWhy every developer should be a data scientistIvan Danyliuk · GDG DevFest Ukraine 2017
• 13mEnterprise-Level Data Science SuccessPrabhakar Thanikasalam · rstudio::conf(2022)
• 44mPeople, Technology, Culture: How to make Data Science workMario Meir-Huber · Data Science Conference 4.0