Upload
apigee
View
1.337
Download
0
Embed Size (px)
DESCRIPTION
Watch the video: http://youtu.be/KFLdWjN0n_k Customer expectations for relevant and individualized experiences are rising and evolving at breakneck speed. This has enterprises working furiously at building data infrastructure to collect and store data. But collecting and storing is only the beginning. The technology and know-how to derive value from data—to do predictive analytics on big data—is fast becoming the critical competitive differentiator for businesses. Join Apigee’s Abhi Rele and Alan Ho as they discuss the market dynamics of predictive analytics and big data and the key capabilities needed to deliver the adaptive apps and APIs every business needs to remain relevant and be competitive. Join to Discuss: - Data lakes, machine learning, unstructured data processors, real-time access, APIs—the capabilities to rapidly deliver predictive analytics on big data - Getting from data lake to production app - how putting big data to use and deriving real value requires a fresh approach - Pros and cons for the build vs. buy decision to deliver adaptive apps and APIs
Citation preview
Predictive Analytics on Big Data
DIY or BUY?
@karlunhoAlan Ho
@abhireleAbhi Rele
youtube.com/apigee
slideshare.com/apigee
www.iloveapis2014.com
Use BIGDATA10 for 10% off
Agenda
• Predictive analytics on big data
• Businesses are conflicted
• Forging a path forward
CC-BY-SA
Why predictive analytics on big
data?
CC-BY-SA
The new normal
• Omni-channel
• Individualized
• Proactive
CC-BY-SA
Challenges
• Data lakes: learning to swim
• Predictive analytics: in flux
• Open source: rapid innovation
• Got data scientists?
• Point solutions
CC-BY-SA
Key conflict
DIY with open source
OR
BUY product
CC-BY-SA
Evaluating options
CC-BY-SA
DIY BUY
Pros• Control
• Cost savings
• Time to market
• Market
evolution
Cons• Expertise
• Risk
• Hype
CC-BY-SA
Data lake
Descriptive analytics
Predictive analytics
Integration
Mo
nito
ring
& m
gm
t.Mobile Web Kiosk IoT
Unstructured & structured data
Event & entity data
Real-time & batch data
Partner
Internal & external data
Data lake
• Hadoop
• Entities and events
CC-BY-SA
Data lake
Descriptive analytics
Predictive analytics
Integration
Mon
itori
ng
&
mg
mt.
Descriptive analytics• Simple
• Complex
CC-BY-SA
Data lake
Descriptive analytics
Predictive analytics
Integration
Mon
itori
ng
&
mg
mt.
Predictive analytics
• Summarized vs. fine-grain data
• Unstructured data
• No open source winner
• Difficult to use
• Mahout vs. Oryx vs. RHadoop
CC-BY-SA
Data lake
Descriptive analytics
Predictive analytics
Integration
Mon
itori
ng
&
mg
mt.
Integration• APIs vs. useful APIs
• Real time
• Scalability
• Security
CC-BY-SA
Data lake
Descriptive analytics
Predictive analytics
Integration
Mon
itori
ng
&
mg
mt.
Monitoring & mgmt.• Achilles heel
• Model performance
• Model deployment
• Availability
CC-BY-SA
Data lake
Descriptive analytics
Predictive analytics
Integration
Mon
itori
ng
&
mg
mt.
to summarize…
DIY or BUY?
CC-BY-SA
CC-BY-SA
Data lake
Descriptive analytics
Predictive analytics
Integration
Mo
nito
ring
& m
gm
t.Mobile Web Kiosk IoT
Unstructured & structured data
Event & entity data
Real-time & batch data
Partner
Internal & external data
DIY considerations
• Maturity of open source
• Skills and expertise
• Ability to execute
• TCO
CC-BY-SA
BUY considerations
• Hype vs. reality
• Time to market
• Control & flexibility
• True ROI
CC-BY-SA
www.iloveapis2014.com
Use BIGDATA10 for 10% off
Questions?
@karlunhoAlan Ho
@abhireleAbhi Rele
Thank you!