Upload
mark-tabladillo
View
295
Download
0
Embed Size (px)
DESCRIPTION
Many companies are starting or expanding their use of data mining and machine learning. This presentation covers seven practical ideas for encouraging advanced analytics in your organization.
Citation preview
7 Ideas on Encouraging Advanced Analytics Mark Tabladillo Ph.D.
Microsoft MVP, SAS Expert; Trainer & Consultant, SolidQ
July 17, 2014
AbstractMany companies are starting or expanding their use of data mining and machine learning. This presentation covers seven practical ideas for encouraging advanced analytics in your organization.
MarkTab
Microsoft MVPSAS Expert
Trainer & ConsultantData Scientist
Associate Faculty – University of Phoenix (School of Advanced Studies)
@marktabnetLinked In
http://marktab.net
PremiseAdvanced Analytics promises to handle the common challenges facing organizations
How do we respond to: Volume, Velocity, Variety
How do we achieve rapid analytics
How do we develop technology
How do we obtain more skilled analysts and data scientists
How do we tell stories
Scientific MethodBaseline = Null Hypothesis
Alternative = Alternative Hypothesis
Questions:
Is there evidence to reject the null hypothesis?
How do you know that?
So what?
Epistemology: Science relies on presuppositions
Seven AreasAdvanced Analytics promises to handle the common challenges facing organizations
1-3 How do we respond to: Volume, Velocity, Variety
4 How do we achieve rapid analytics
5 How do we develop technology
6 How do we obtain more skilled analysts and data scientists
7 How do we tell stories
VolumeBaseline: Ignore it
Alternative:
Technology (flat files, tape, CSV Hadoop)
Sampling
VelocityBaseline: Ignore it
Alternative
Streaming (StreamInsight)
Sampling
VarietyBaseline: Ignore it
Alternative:
Different database typesSQL
NoSQL: Excel, Power Pivot, OLAP, Graph, HDInsight, Hadoop
Sampling
Achieving Rapid AnalyticsBaseline: IT (Information Technology) produces, business units consume
Alternative:
Business Units share production and consumption with Information Technology
Approach: Learn the business, work on better data
Developing TechnologyBaseline: Let the vendors do it
Alternative
Build it
Virtual Machines: Cloud, On Premise, Hybrid
Development Environments D = Development
R&D = Research and Development
Obtaining TalentBaseline: Ignore the issue
Alternatives
Buy
Rent
Create
Lead
StoriesBaseline 1: internal focus because we’re just like everyone else
Baseline 2: the whole world is unique with no unifying patterns
Alternative
Technology conferences
Industry conferences
Benchmarking
Eric Siegel: Predictive Analytics
One Book to ReadThomas Kuhn: The Structure of Scientific Revolutions
Tommy Lasorda, Manager LA DodgersYou can make it happen
You can let it happen
Or you can wonder, what happened?