Some emerging trends in analytics

Some Emerging Trends in Analytics
Prescriptive Analytics
In-memory Computing

3/1/2014


1

In God we trust, others must bring Data

-W Edwards Deming
3/1/2014


2

Prescriptive Analytics

Introduction, Origin & Process, Applications

3/1/2014


3

Introduction
(not just what and when, but also why…)





It automatically synthesizes big data, multiple disciplines
of mathematical sciences and computational sciences,
and business rules, to make predictions and then suggests
decision options to take advantage of the predictions.
It can continually take in new data to re-predict and represcribe, thus automatically improving prediction accuracy and
prescribing better decision options.

3/1/2014


4

The Final Step in Analytics…
Descriptive

• The use of data to find
out what happened in
the past
• Data modeling, trend
reporting, regression
analysis

Diagnostic

• The use of data to find
out why it happened
• Postmortem analysis,
trend analysis

Predictive

• The use of date to find
out what could happen
in the future
• Data mining, predictive
modeling

The Analytics
Sequence

Prescriptive

3/1/2014

• The use of data to
prescribe the best
course of action for the
future
• Optimization, simulation


5

Why is It Needed?
 Prescriptive model emerged to check the shortcomings of the
predictive models.
 Optimal solution is not realistic, so prescriptive models predict
case-based suggestions
 Integration of the descriptive and predictive components is
critical
 Availability of huge data and advanced tools has made
application of this model more feasible

3/1/2014


6

If you don’t know how to ask questions,
you know nothing.

-W Edwards Deming
3/1/2014


7

Origin of Prescriptive Analysis
 Got its early start in 2003 with the R&D focussed organisation
DataInfoCom which was later renamed as AYATA.
 During 2004 and 2005, AYATA conceptualized the technology
footprint for Prescriptive Analytics and went on to release first
version of the software in 2007.
 In 2011 it was trademarked for two categories: Software and
Software as a Service.
 AYATA has worked with Apache Corp. to improve production by
proactively reducing pump failures in the field.
 Later deployed at various Fortune 100 companies including
Dell, Microsoft & Cisco.
3/1/2014


8

Prescriptive Analytics Process
Identify
objectives
(values)

More analysis
needed?

Identify
Alternatives

Perform
Sensitivity
analysis

Decompose &
Model
Problem

Choose best
alternative

3/1/2014


If NO,
Make
recommendation

9

What gets measured, get managed.

-Peter Drucker
3/1/2014


10

The Prescriptive Process Analyses

Potential
decisions

3/1/2014

The interactions
between
decisions


The influences
that bear upon
these decisions

11

Applications
Health care
strategic
planning

• To leverage operational and usage data by combining the
external factors such as economic data, population
demographic trends and population health trends
• This helps us to more accurately plan for future capital
investments such as new facilities and equipment utilization,
understand the trade-offs between adding additional beds,
expanding an existing facility versus building a new one.

Oil and gas
industry

• Prices fluctuate dramatically depending upon supply, demand,
geo-politics, & weather conditions.
• Gas producers have a keen interest in more accurately
predicting gas prices so that they can lock in favourable terms
while hedging downside risk.
• Prescriptive analytics can accurately predict prices by modelling
internal and external variables simultaneously and also provide
decision options and show the impact of each decision option.

3/1/2014


12

Applications

Telecommunications
and cable

3/1/2014

• Telecom and cable companies have large field service
organizations to install, repair and resolve issues at
customer sites, both in homes or at businesses.
• Mastering field service operations is critical because
field services impact customer satisfaction.
• Prescriptive analytics enables telecom and cable
providers to dramatically improve effectiveness and
efficiency of field service operations.


13

Those who do not remember the past,
are condemned to repeat it.

-George Santayana
3/1/2014


14

In-memory Computing

Introduction, Benefits, SAP HANA, Applications

3/1/2014


15

In–Memory Computing
In-memory computing is the storage of information in the main random access
memory (RAM) of dedicated servers rather than in complicated and
comparatively slow disk drives.

Source: SAP AG

3/1/2014


16

Why In-Memory Computing
 Ground breaking Innovation
 10,000x improvement in speed of access

 Movement to main memory from disk storage:
 Viable performance with increasing data volumes
 Affordable servers > 1 TB system memory
 CPUs – Multi core for rapid parallel processing
 Cost feasible technology
3/1/2014


17

The price of light is greater than the
cost of darkness.

-Arthur C Neilsen
3/1/2014


18

Benefits of In-memory Computing
 ‘In-memory computing’ is a software that could transform
business IT enabling companies to crunch and analyse large
volumes of data in near real-time and run sophisticated ‘what
if’ simulations.
 In-memory computing uses sophisticated data compression
techniques to store information in Ram which is 10,000 times
faster than standard disks, enabling companies to analyse that
data in seconds instead of hours.
 Empowerment and increased flexibility is assured because it
reduces business users reliance on IT.

3/1/2014


19

 Cost benefits are assured because it reduces hardware and
maintenance costs through a flexible, cost-effective, real time
approach for managing large data volumes.
 Real time visibility is assured as data is pushed from the
sources to the warehouses.
 Previously the sheer volume of data and computational power
allowed only for pre- determined data analysis. With inmemory systems detailed data is loaded into memory where
calculations are performed “ on the fly” at query time.

3/1/2014


20

Not everything that can be counted
counts, and not everything that counts
can be counted.

-Albert Einstein
3/1/2014


21

SAP HANA: (High Performance Analytic Appliance)





Engine of Real-time enterprise
Provides foundation to build new generation of Analytics applications
Enables customers to analyze data from virtually any source, in real time
Example showing actual customer performance of a core reporting process.

Runs 350x
faster
13 seconds
77 minutes

With SAP HANA

Before SAP HANA

Source: SAP AG

3/1/2014


22

HANA Accelerates Data,
Innovated with AnalyticsApplications, Analytics
is the Database LOB
ERP &
Database
Mobility
Mobility
Accessible Systems

Data “In”

ERP + LOB

BICS

Systems of Record

Business
Analytics

Info “Out”

Systems of Engagement

Business Applications Performance
Bound by Data

Oracle,
Oracle
DB2

ELT or ETL

DB2

HANA
In Memory Database

ELT or
SQL ETL

SQL,
Other
Other
Source: SAP AG

3/1/2014


23

In-memory Computing
Traditional Computing using HANA
as Data Layer
Local BI
Data
Mart
HANA
DB

SAP ERP 1
Database

Local BI
SAP
ERP 1

SAP
Data
ERP 2
Mart
HANA
DB

Corporate BI

Corporate BI

NEW
NEW
SAP ERP 2
APP 1
APP 2

Enterprise Data
Enterprise Data
Warehouse (BW)
Warehouse (BW)

Database

HANA

Database
HANA
BWA

Local BI
Data
Mart
HANA
DB

NON SAP

NON
SAP

Database
Source: SAP AG

3/1/2014


24

It is a capital mistake to theorize before
one has data.

-Arthur Conan Doyle
3/1/2014


25

Applications of In-memory Computing
 Research shows that organizations that have adopted inmemory computing were not only able to analyze larger
amounts of data in less time than their competitors – they
were literally orders of magnitude faster.

3/1/2014


26

 Two birds with one stone : Volume and Velocity
 Relies on latest breed of high powered servers
 Vast amount of RAM where business information can be
directly stored
 Zero latency
 No look up time
 High speed
 Multi-core processing
 This results in real-time analysis of core business data

3/1/2014


27

 It is most applicable for large enterprises
 Steep initial investment
 43% of companies using in-memory computing have an
annual revenue of over $1 Billion
 Data should be available in structured format(transaction
data, sales figures or product codes)
 Should use appropriate analytical applications
 When it comes to processing large amounts of structured data
very quickly, in-memory computing outperform their peers
 Almost 3 times the data over a hundred times the speed
 Data volumes are growing at the rate of 36% per year
3/1/2014


28

3/1/2014


29

Thank You!

Srushti, Rahul, Saurabh, Achintya, Divya, Prashant, Krithika

3/1/2014


30

Some emerging trends in analytics

More Related Content

What's hot (20)

Similar to Some emerging trends in analytics (20)

More from Prasant Patro (20)

Recently uploaded (20)

Some emerging trends in analytics