predictive analytics for the business analystdownload.101com.com/pub/tdwi/files/actuate...

Post on 04-Oct-2020

10 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Predictive Analytics for the Business Analyst

Fern Halper

July 8, 2014

@fhalper

Sponsor

3

Speakers

Fern Halper Research Director for

Advanced Analytics,

TDWI

Allen Bonde VP, Product Marketing

and Innovation,

Actuate

• Brief overview of predictive analytics

• Predictive analytics trends

• Skills required

• Getting started

Agenda

4

The Analytics Spectrum

5

Excel

Dashboards and reports

Other BI

Visualization

Advanced Analytics

Often becomes more algorithmic

Predictive Analytics

6

A statistical or data mining solution

consisting of algorithms

and techniques that can be used

on both structured

and unstructured data to determine outcomes.

• Marketing and sales**

• Healthcare

• Fraud detection

• Human resources

• Operations maintenance

• And many, many more!

Some Use Cases

Popular Methods

• Decision Trees

• Regression

• Cluster Analysis

8

Total

Monthly bill

> 2 yr

>$100

Length of time

customer

Call center

calls No

< 2 yr

<$100

No. of phones

on account

90% probability

Etc…

Trends

• Ease of use

• Disparate data types

• Operationalizing and

embedding advanced analytics

• Prescriptive

9

1. Ease of Use

• Graphical UI

• Automation

• Solutions

10

Ease of Use

Predictive

Analytics

New builders

11

New Users are Emerging

Statistician/Modeler Moving towards critical

thinker with knowledge of

the business -- e.g. a

business analyst

The Business Analyst Rules

13

(source: TDWI Best Practices Report Predictive Analytics for Business Advantage, 2014)

With an Evolving Skill Set

14

This ranks low

2. Data, Data

Predictive

Analytics

New Data

Types

15

New and “Big” Data Types for Analysis

16

Disparate

data sources

3. New Deployment Models

Predictive

Analytics

Operationalizing

Embedding

17

Embedded Analytics

18

An example:

Example: Fraud

• Objectives:

– Reduce fraud

– Improve customer experience

• Benefits

– Speed up process

– Reduce false alerts

– Save money

19

4. Prescriptive Analytics

Whereas predictive analytics helps to determine

what might happen, prescriptive analytics takes

this further to either suggest or automatically

initiate a subsequent action based on this

output.

20

Skills Needed (1)

21

Framing the problem

2. Data Sense

3. Domain

Expertise

1. Critical Thinking

1. Critical Thinking

• Ability to formulate a question

• Comfortable creatively thinking in numbers

and attributes

• Interpretation skills

• Inference

Above all: Questioning

22

2. Domain Expertise

• Helps in:

– formulating good questions

– understanding objectives

– assessing the model and taking action on it

• Understanding relevant data

– Dealing with data – outliers, missing data, etc.

23

3. Understanding data

• Target vs. explanatory variables

• Derived variables

• Lots of new data types

– Documents, graph, location

– May require parsing, geocoding

24

Skills Needed (2)

25

Explain/Defend

5. Techniques

4. Tools

4. Understanding the tools!

26

5. Understanding the techniques

• A basic understanding is necessary

27

6. Storytelling

• Don’t start with the techniques

• Begin with the business problem and the

outcome.

28

(source: vitualspeechcoach.com)

Getting started

• An analytics program does take time

• But you can get started quickly

29

Not necessarily sequential!

Getting started

• Pick a problem

• Experiment and involve business/IT

• POV/POC tied to metrics

• Decide beforehand how to integrate it into

a process

• Balance cost of model with model building

solutions

30

31

Actuate Corporation © 2014

Fast is the new Big!

32

“Fast analytics” enables…

Iterative process employ A/B testing, chunk down problem

Better questions “what do I ask next?”

Instant feedback “what can I adjust?”

33

Actuate Corporation © 2014 34

Staying Focused is Key

Improved cross-sell?

Getting all data together

Pricing optimization

What’s the business case?

Better customer understanding

Risk management

35

Actuate Corporation © 2014 36

Picking the Right Tool

Key Challenges…

Disparate sources, billions of records

Complexity of loading, cleaning

Need all data in one view

Easily profile and segment

Look for trends, relationships

37

Access data Understand Patterns

Deliver Insights

Explore, visualize…with no

coding!

Enable non-technical business users

Support iterative, collaborative work

Integrate with operational systems

Insights in minutes vs. days

CRM/ERP

Web

Social

“To create a complete picture of customers, we need to combine insights from social channels and campaigns with Web and transactional data”

10 Reasons 2014 will be the Year of Small Data, ZDNet, Dec 2013

Other sources

Columnar

DB

APPROACH: Make it easy to access and integrate data…

38

Exploration &

Visualization

“Many vendors are trying to (make predictive analytics available to an end user in a consumable form) but in our view BIRT Analytics comes closest to getting it right, by …

not requiring the user to select algorithms” IDC, Feb 2013

Profile

Forecast

Decision Tree

Cluster

Analytics &

Data Mining

Segmentation

Campaign

Workflow

Columnar

DB

…and shorten time-to-value by using pre-built analytics

39

40

visit www.actuate.com/BIRTanalytics

my blog: www.smalldatagroup.com follow @actuate, @abonde

42

Questions?

43

Contact Information

If you have further questions or comments:

Fern Halper, TDWI fhalper@tdwi.org

Allen Bonde, Actuate abonde@actuate.com

top related