Four Emitters Connected to the Internet of Things - BuckeyeDAMA

youthfulgleekingRéseaux et Communications

17 févr. 2014 (il y a 3 années et 1 mois)

121 vue(s)

Big Data

What does it
mean
?

How do we
mine

it?


John Johansen

October 2013

jjohansen@agiletech.com


Session Objective

Who saw

that

coming!



Our organizations’ data is increasingly capable of helping us anticipate
and
plan
for the
unexpected.




Powerful

tools are emerging to help

identify patterns and make predictions of potential

risks and opportunities.


These

predictive analytics

will allow

companies

to focus

on the real trouble spots and develop the right conclusions.



This
session will explore and demonstrate these tools while identifying potential applications

and
solutions

in our day
-
to
-
day jobs.





At
the conclusion of the session participants will understand


The role that analytics can play in
supporting your organizational objectives.


How these tools can identify elevated risk and help plan effective
strategies to maximize opportunities.


Who
saw that coming?

You did
!

2

Agenda


Big Data Explained



Why Are These Solutions Emerging Now ?



Some
Common and Not
-
so Common Applications in our Businesses



The Steps in the Mining Process



A Real Live Demo of a Mining Process



Questions



Wrap
-

Up


3

But first… a
promise


4

But first… a
promise


5

But first… a
promise


6

No plunge !

Big Data Explained

7

Volume

Velocity

Variety

RF Tagging

Healthcare Monitoring

Unstructured Data

Sensor Data

Real Time Data

>
Peta
?
Zetta

Voice of the Customer

Text Analysis

Four Emitters Connected to the Internet of Things


Cows & Crops



8

Transactions, Interactions & Observations

9

Big Data in the Gartner Hype Cycle

10

There are less scientific hype indicators

11

Why Now?


Chances are, the tools you need are the tools you have.*



At long last, the data that you need is the data that you have.



The processing power that you need is the processing power that you
have.*


12

* And if you don’t already have them, they are readily available
with very reasonable
ROI

Four Emitters Connected to the
Internet of Things


Cows & Crops



Christmas Trees



13

Data is Increasingly a Differentiator

14

Production Reporting

Interactive Dashboards

Data Discovery

Predictive Analytics

Operational Analytics

Differentiation

Sophistication

High

Low

Low

High

Predictive Analytics

15

Potential Challenges

16

“Prediction is very difficult…



Niels

Bohr

Casey Stengel

President Bill Clinton

Especially if it’s about the future.”

Predictive Analytic Solutions


We see these solutions as a class of software tools that will look
through large sets of data to uncover subtle patterns in data, then use
those patterns to predict the behavior of a new set of data.



The
impact these solutions are having at companies are generally in
the areas of:


Improving the profitability of existing clients by identifying high probability
cross
-
selling activities.


Improving the effectiveness of direct marketing programs through more
focused, higher likelihood of success programs.


Identifying fraud.


Identifying likely candidates for churn either in the customer
-
base or in the
channel.


Modeling customer reactions to price or term changes
.

17

Organizations Embrace Analytics Differently

Killer Applications

What are the steps?


As with any other solutions, we suggest starting at the end: Outlining
specifically what we are hoping to achieve.



Next we need to establish at least two sets of data from our existing
historical data. One will help us
Train
the model, the other will be a
Test

set that will allow us to validate that in fact the model that we create is
valid.



Once we’re sure we’ve got a model that establishes the right
relationships, we can run our source data through our model and get our
predictions.



Analyze, Rinse. Repeat
.

20

Four Emitters Connected to the
Internet of Things


Cows & Crops



Christmas Trees



Pill Bottles





21

Demonstrating the Power


We’ve arranged for a quick demo of the Microsoft tools.



Churn Prediction



Claim Amount



Cross Selling



It’s interesting to note, that this demonstration is running on a simple
server, using software tools that are included, free, with Microsoft
SQLServer
.


22

Demonstration


Predicting Churn


23

Demonstration


Claim Cost


24

Demonstration


Cross Selling


25

Four Emitters Connected to the
Internet of Things


Cows & Crops



Christmas Trees



Pill Bottles



Diapers



26

Questions?

27

In conclusion…


Big Data Explained



Why Are These Solutions Emerging Now ?



Some
Common and Not
-
so Common Applications in our Businesses



The Steps in the Mining Process



A Real Live Demo of a Mining Process



Questions

28

In conclusion…

29

“I figure lots of predictions is best.
People will forget the ones I get
wrong and marvel over the rest.”

Alan Cox

Contact

John Johansen

Partner

Agile Technologies, LLC

One Easton Oval, Suite 388

Columbus,
Ohio
43219


jjohansen@agiletech.com


www.agiletech.com


30