12. Analytics
From Traditional To Big Data
BI & Analytics Big Data Analytics
Infrastructure
Traditional Data
Big Data Management
Management
(Hadoop & NoSQL)
(RDMS & EDW)
Smaller;Traditional Bigger;Contemporary;
13. Analytics
Traditional Data Professionals
Data Analysts BI Team
BI & Analytics
Data
Developers DBAs Warehouse
Engineers
Infrastructure
Traditional Data
Management
IT or Ops
(RDMS & EDW)
Smaller;Traditional
14. Big Data Professionals
Data Scientists
Analytics
Data Analysts Business Analysts
Big Data Analytics
Data Developers
Infrastructure
Big Data
Management
IT or Cloud Ops
(Hadoop & NoSQL)
Bigger;Contemporary;
15. Empowering Analytics Teams
Data Scientists
Data Analysts Business Analysts
●Do Effective Groundwork
・Reach out for training and services
・Collaboration is King
・Don t silo the data or the knowledge
●Involve Analytics team from the start
・Empower your data scientists, data and business analysts
with purpose-built strategic tooling
・ Use the cloud for ramping and, if possible, production
・ Don t route everything through IT. This isn t Grandma s BI.
20. 優れた機械学習アルゴリズムの条件
・Must have parameter for dialing complexity up and
down
・fast to train
・doesn't need entire data set in memory at one time
・MRable
・easily implementation
・handles numeric
21. 最近、注目されているアルゴリズム
Glmnet
one of the best algorithms in machine learning today.
background - ordinary least squares regression,
limitations of OLS