SLIDE 1
Where We Are Today
2
The Modern Analytical Landscape 2 Where We Are Today 3 Data - - PowerPoint PPT Presentation
The Modern Analytical Landscape 2 Where We Are Today 3 Data - - PowerPoint PPT Presentation
The Modern Analytical Landscape 2 Where We Are Today 3 Data Explosive growth Data 2014 566PB/day 2017 est 1.5EB/day 2020 est 2.5YB/day 4 Data Types Telemetry IoT Formats JSON/BSON
SLIDE 2
SLIDE 3
Data
▸ Explosive growth
▹ Data ▹ 2014 – 566PB/day ▹ 2017 – est 1.5EB/day ▹ 2020 – est 2.5YB/day
3
SLIDE 4
Data Types
▸Telemetry – IoT ▸Formats
▹JSON/BSON ▹Audio/Video ▹CDISC ▹F004 ▹ASDF ▹DMS-MTX
4
SLIDE 5
Usage
▸Analytics recognition ”I keep saying the sexy job in the next ten years will be statisticians.”
Hal Varian – Google Chief Economist 2009
5
Finally.
SLIDE 6
6
SLIDE 7
7
SLIDE 8
Used to be Easier
▸Minimal Sources ▸Minimal Data types ▸Smallish Volume ▸Couple of Analytic Packages
8
SLIDE 9
9
SLIDE 10
10
SLIDE 11
Used to be Harder
▸Minimal Sources ▸Minimal Data types ▸Smallish Volume ▸Couple of Analytic Packages ▸Finding data ▸Gathering data ▸Limited data types ▸Limited Analytic Techniques ▸Limited Compute Power
11
SLIDE 12
Current “Trends”
Disruptive Technologies
12
SLIDE 13
Hadoop
▸“Released” in 2006 building on work done in 2003/4 ▸Developed for page rank ▸Distributed File Storage ▸Map-Reduce on top ▸Apache Hadoop 1.0 released in 2012
13
SLIDE 14
Hadoop
▸Ideal for particular use cases ▸Not always performant for Analytics
14
SLIDE 15
Data Science
▸Gen 1 Data Scientist
▹Programmer first
▸Gen 2 Data Scientist
▹Taught at University ▹Has stats knowledge
▸Gen 3 Data Scientist
▹Workplace experience
15
SLIDE 16
Data Science
▸ Becoming the new “I.T department” style roadblock ▸ Rapidly becoming a 4 letter word
16
SLIDE 17
Machine Learning
▸Trendy term ▸Highly sought after skills
17
SLIDE 18
18
Congra gratul ulation ions y you c can now put ut m mach achin ine le learnin arning o
- n y
n your ur resumes.
SLIDE 19
Deep Learning
▸Simply a neural network with >1 hidden layer ▸Frameworks
▹Tensorflow ▹Theano ▹Keras ▹Caffe ▹DSSTNE t
19
SLIDE 20
Deep Learning
▸May produce superior results ▸Only way (currently) for some problems ▸Programming skills ▸Over complicating things ▸Loss of explainabilty
20
SLIDE 21
In Database
▸Move the work to the data ▸Database to manipulate the data ▸SAS and Teradata 10 year partnership ▸Phenomenal reductions in processing
21
SLIDE 22
In Memory ▸SAS leading the way ▸Visual Analytics/Statistics ▸HPA procs ▸Teradata 750 appliance ▸Viya
22
SLIDE 23
23
THANKS!
Any questions? You can find me at @pwsegal-ca