Top Banner
#analyticsx Copyright © 2016, SAS Institute Inc. All rights reserved. Prescriptive Analytics Prescribing Faster and More Accurate Decisions
22

Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

Jun 11, 2020

Download

Documents

dariahiddleston
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Prescriptive Analytics

Prescribing Faster and More Accurate Decisions

Page 2: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Agenda

• Value of prescriptive analytics

• Building prescriptive analytics• The business of business rules

• Analytic inventories

• In database performance gains

• Deployment and operationalization

• Prescriptive analytics for the Internet of Things (IoT)

Page 3: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Why do analytics?

“takes data not guesswork”

James Taylor, IIA presentation, June 23rd, 2016

IIA poll, June 23rd, 2016

Page 4: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Intent without execution

James Taylor, IIA presentation, June 23rd, 2016

Page 5: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Prescriptive analytics drive value that is…

• Consistent – explicit, traceable• Accurate – valid, flexible• Relevant – managed, timely• Scalable – independent, portable

Page 6: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Specific Big Data Analytics Challenges

• Model Scoring takes significantly too long

• We spend too much time doing data prep

• Data discover is taking too long because of the volume of data

• Prioritizing which groups to support

Page 7: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Value of Prescriptive Analytics

Page 8: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Decision Agility

Page 9: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Building Prescriptive Analytics

Page 10: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Business Rules

Relationships• Simple Rule Flow

• Construct Rules as Rule Sets

• Rule Sets Ordered in Rule Flows

• Rule Flows Published and Managed

Rule 2Rule 3…

Rule Set A

Rule 1Rule 2

Rule 3…

Rule Set B

Rule Flow

Rule 1Rule 2

Rule 3…

Rule Set C…

Vocabulary(Entities,Terms)

A Rule Flow is a

collection of Rule Sets. Execution is defined by

order of Rule Sets.

Rule Set action values are available in

subsequent Rule Sets in the same Rule Flow.

Rule 1

Rule 1Rule 2

Rule 3…

Rule Set B

Page 11: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

SAS® Business Rules Manager• Support across lifecycle

• Rule Authoring/Discovery

• Rule Testing• Design, Test, Design some more..

• Rule Publishing• Teradata, Hadoop, …

• Services

Vocabulary

Rule

Management

Business

Rules Repository

Rule Flow

Publishing

In Database + Hadoop

Rule Author

Rule

Flow Testing

Page 12: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Why formally manage business rules?

Formalize business processes

Rapid development & deployment

Governance & compliance

Reuse assets & deploy anywhere

Capture in-depth testing

Increase business agility

Secure authorization

Page 13: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Analytic Inventories: SAS® Model Manager

Page 14: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

!!!???!!!

The ‘IT’ folks The ‘Analytics’ folks

I just built 850 new models. When can you

put them into production?

Page 15: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Building Prescriptive Analytics

Decisions for:- Batch processes- Web services

Page 16: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Results from Prescriptive Analytics

• Consistent to scenarios

• Analytically sound

• Traceable

• Designed for automation

Streamline deployment

Increase productivity

Increase Confidence

Value and efficiency for the organization

Page 18: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

• Internet of People is mainstream

• In < 20 years, we’ll each communicate with 5,000 ‘things’1

• B2B uses could generate up to 70% of potential value enabled by IoT2

• More than 5.5 million new things are connected every day

1 As projected by Dr. John Barrett, Head of Academic Studies, Cork Institute of Technology, Oct. 20122 M. Mendoza, $11 Trillion: Potential Economic Impact Of Internet Of Things By 2025, TechTimes, July, 2015

The IoT

Page 19: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Prescriptive Analytics for the IoT

AnalyticsDATAB I G

IOT High Velocity

Complex

Large

New Business Models

Quality of Life

Early Warnings

Efficiencies

New

Value

Act

Understand ActSense

Page 20: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Questions?

Page 21: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

#analyticsx

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

Conclusions• Be more responsive, proactive and reliant on data-

driven operational decisions for new opportunities.

• Improve performance and minimize time previously spent moving or duplicating data and code between systems.

• Increase security and compliance of data in one integrated, highly governed environment.

• Be consistent & automate

Page 22: Prescriptive Analytics - SAS · Specific Big Data Analytics Challenges •Model Scoring takes significantly too long •We spend too much time doing data prep •Data discover is

C o p y r ig ht © 201 6, SAS In st i tute In c. A l l r ig hts r ese rve d.

#analyticsx