Date post: | 18-May-2015 |
Category: |
Technology |
Upload: | arago-gmbh |
View: | 1,707 times |
Download: | 0 times |
Particle CollisionIn Code Space
API meets Big Data
Chris Boos (@boosc)[email protected]
How to Web, Bucarest 8.11.2012
Mittwoch, 14. November 12
Big Data?
Mittwoch, 14. November 12
Data, lots of itMittwoch, 14. November 12
Even in simple datasets, common statistics fails - (avg, min, max, distribution)
Mittwoch, 14. November 12
Wisdome2010s Systems Thinking
Intelligence2000s Knowledge Ecology
Knowledge1990s Knowledge Management
Why you need big data
Yield
Data1960s1950s Data Processing
You Are Here !
Information1980s1970s Information Management
Mittwoch, 14. November 12
Because, (now) We Can!
Mittwoch, 14. November 12
79 times more CPU power than used in Apollo missions on one iPhone 4
Mittwoch, 14. November 12
All the resources you like at your fingertips
Mittwoch, 14. November 12
Buzzword Bingo
Big DataData Engineer
H-Space
HadoopCassandra HBasePIGredis.io Eucalyptus
Machine Learning Support Vector Machines
Gaussian ProcessesSwarm Intelligence
Genetic Algorithms
Agents/Bots
R+Natural Language Processing
ClusteringCore Dataset
NoStats
Mittwoch, 14. November 12
Data Exploration Is a Big Adventure
Mittwoch, 14. November 12
Big Data is all about improving prognostic power - guess what it is NOT!
What people think
Mittwoch, 14. November 12
Finding clusters, evaluating outliers and interpreting white noise
Warming up
Mittwoch, 14. November 12
You are not looking for patterns, you are looking for anomalies
Mittwoch, 14. November 12
Two tips for looking at data:
1. Plot it
2. Remove all labels
Mittwoch, 14. November 12
3 Ways to solve a coding or
data problemElegant
Standard
Brute Force
Mittwoch, 14. November 12
The Sledge Hammer of Big Data - Map Reduce
Mittwoch, 14. November 12
Old Style (Imperative) Programming
• Step by step explanation what to do
• Explaining WHAT to do rather than RESULTS you want
• Always necessary for basic algorithms
1
2
3
Mittwoch, 14. November 12
One New Style (Functional) Programming I
• Combine results to become a program
• Allows dynamic distribution
• Map-Reduce is only one way of doing it!
1
2
3
Mittwoch, 14. November 12
Functional Programming II
F ( G ( H ( A,B) , C), D)
getMusicLikes(getFriends(facebookID)
Instead of
for i in getFriends(facebookID) getMusicLikes(i)
Mittwoch, 14. November 12
BASE(Basically Available, Soft State, Eventual consistency)
not
ACID(Atomicity, Consistency, Isolation, Durability)
Mittwoch, 14. November 12
That is all batch processing!?
Mittwoch, 14. November 12
The big in memory trend is delaying the problem
Mittwoch, 14. November 12
Google is going well beyond Map Reduce and batch processing
Mittwoch, 14. November 12
And at facebook they are going far beyond distributed data storage and processing
Mittwoch, 14. November 12
Most people using Big Data end up building their own platform
Mittwoch, 14. November 12
APIs, for external use only?
NO WAY!
Mittwoch, 14. November 12
Monolithic architectures are so out!!!
Mittwoch, 14. November 12
Do not use Design Patterns as an excuse to get lost in OO space!!
Mittwoch, 14. November 12
Standardising dev tools can mean 2 things:
You are a little dictator and could not find a country
Your team is not as good as you say and needs guidance
Mittwoch, 14. November 12
API 1.0
Mittwoch, 14. November 12
Function calls or servicesMittwoch, 14. November 12
Stop religion on transfer format (XML, JSON), who cares?
Mittwoch, 14. November 12
Stop religion on service design (SOAP, REST), be practical!
Mittwoch, 14. November 12
API 2.0
Mittwoch, 14. November 12
Adding semantics means adding context
Mittwoch, 14. November 12
OWL or the theoretic vision of everyone speaking the same language
and meaning the same thing
Mittwoch, 14. November 12
Facebook‘s pragmatic approach: the open graph, sort of RDF
Mittwoch, 14. November 12
API Management and Monetization
Mittwoch, 14. November 12
oAuth vs. application tokens - or both?
Mittwoch, 14. November 12
Monetization models
API calls
Data transfer
Data access
Mittwoch, 14. November 12
Not necessarily DIY, why not use an external API manager?
Mittwoch, 14. November 12
CollisionBIG Data and APIs
Mittwoch, 14. November 12
Your platform cannot handle infinite exploration
Mittwoch, 14. November 12
Your users will not understand the mechanics behind your data processing
Mittwoch, 14. November 12
Your data has to be put into context
Mittwoch, 14. November 12
Then you can onboard new team members quickly and get external developers to use your stuff
Mittwoch, 14. November 12
Check out my tool list:http://www.hcboos.net/100-links/
Mittwoch, 14. November 12
Credits• Images on pages 8,11,15,21,22,25,29,31,32,33,34,39,40,41,43,44,45,46
are used with kind permission of CERN, are (c) CERN and are used under the following terms of use: http://copyright.web.cern.ch/
• „Big Data Just Beginning to Explode“ by CSC http://www.csc.com/insights/flxwd/78931-big_data_just_beginning_to_explode
• „Social media network connections among twitter users“ by Marc Smith http://www.flickr.com/photos/marc_smith/
• Asteroid Datasets by Bruce Gary http://brucegary.net/POVENMIRE/x.htm
• iPhone Images (c) Apple Computer inc.
• Google Percolator chart by Marc Bojoly http://blog.octo.com/en/my-reading-of-percolator-architecture-a-google-search-engine-component/
• Prism image by Deviant Art http://sudden2.deviantart.com/art/Floyd-s-Prism-51863247
• Monolith image found on http://www.themonolith.com/
• Design Patterns: Elements of Reusable Object Oriented Software, Erich Gamma, Richard Helm, Ralph Johnson, John Vlissides, 1994, Addison Wesley, ISBN 0-201-63361-2
• OWL Logo (c) World Wide Web Consortium http://www.w3c.org
Mittwoch, 14. November 12
Thank You for Your Time
Chris Boos (@boosc)[email protected]
How to Web, Bucarest 8.11.2012
Mittwoch, 14. November 12
2 Examples
Mittwoch, 14. November 12
The AMP3 Platform at Senzari.comAdaptable Music Parallel Processing Platform
Mittwoch, 14. November 12
MARS-o-Matic at arago.deBig Data based IT modelling and pricing app
More Info at www.mars-o-matic.com
Mittwoch, 14. November 12