The Financial Services Industry Source for Education, Inspiration
DATA - THE SOURCE OF CURIOUS INSPIRATION (AND...
Transcript of DATA - THE SOURCE OF CURIOUS INSPIRATION (AND...
DATA - THE SOURCE OF CURIOUS INSPIRATION
(AND PROGRESS)
Slipstream
15 Apri l 2015
HERE’S AN IDEA: DATA IS THE CREATIVE MEDIUM OF THE
FUTURE! NOW WHERE ARE THE ARTISTS AND WHAT DO
THEY LOOK LIKE?
PURPOSE and MEANING
ROCKSTARDATA
Professor Hans Rosling
Stats that reshape your world view
Western World
Third World
Swedish Students - Karolinska Institutet 1
We
Them
Long life in small family
Short life in large family
1 Hans Rosling - TED Talk 2006
LIBERATE DATA, INSPIRE PEOPLE
Our souls are fed by Curiosity and
Creativity. Data provides a Medium for
achieving this and the tools and methods
already exist for anybody to become the
Artist.
This is BOB
?
Take a
Number!
IT
Speed to Opinion(Weeks/Months)
Opinion-Based Data-Driven
Conventional Self-Reliant
Want
Can
8 %
People
Experience
Time
Want Can
80 %
Speed to Insight(Seconds/Minutes)
Top performing companies:‐ 5x more likely to be data driven‐ Improve productivity by 5 % and profitability by 6 %
Tell a Story.
Start a
conversation.
INSIGHTS - ITS EVERYWHERE AND ITS EASY
BUT UNDERSTAND WHAT THE CREATIVE LANDSCAPE LOOKS LIKE!
DATA IS EVERYWHERE
Social Media
Device Generated
Machine Generated
BIG DATA
UnstructuredData Lakes Internal Systems
Data
Warehouse
Connect
CreateShare
Informal Sources
• http://public.tableau.com/profile/sadethesage#!/vizho
me/CashonHand1/CashonHand
• http://public.tableau.com/s/gallery
LIBERATE DATA , INSPIRE PEOPLE
Imagine if every person had the freedom to transform
data into knowledge easily and accurately.
People need to easily explore, understand, create and
act.
Accessible, data-based insight are the lifeblood of
tomorrow’s organisations and Societies – making them
responsive and innovative.
SPATIALLOCATION PLANNING
8 km
8 km
Aged 18 - 35Income R6.4k to R25.6k
UrbanEmployed
SHOPPING CENTRE
THE ZONE @ ROSEBANK
TARGET DEMOGRAPHIC
YOUNG & ASPIRATIONAL
Data: RSA individual demographics
Data: Shopping Centre (SC) locations
Convert X-Y coordinates to Spatial objects
Find closest SC to each individual, calc proximity
Filter out individuals not within trade area
Filter (split) targeted individuals and others
Count targeted individuals and others in each SC
Ratio of targeted individuals : all
Sort by Ratio, descending
TARGET DEMOGRAPHIC
Young and aspirational: Aged 18 - 35, Income R6.4k to R25.6k, Urban, Employed
Shopping Centre Num. Targeted Individuals Num. All Individuals Percent TargetedBel Air Shopping Mall 2616 14872 17.59Rivonia Crossing 6792 43379 15.66Clearwater Mall 10171 70772 14.37Fourways Crossing 2001 14753 13.56North Riding Square 4051 32173 12.59Monte Casino 2103 17154 12.26Brightwater Commons Mall 4799 40169 11.95Cresta Shopping Centre 5627 56748 9.92Rivonia Square 3329 34543 9.64Cedar Square 3117 35595 8.76Hyde Park Corner 1439 17918 8.03China Mall 8137 101541 8.01The Glen Shopping Centre 10911 142742 7.64Northgate Shopping Centre 5345 71139 7.51Nelson Mandela Square 1035 14022 7.38Greenstone Shopping Centre 10353 143666 7.21Fourways Mall 32 464 6.9Lakeside Mall 9017 140616 6.41Southgate Mall 6530 102756 6.35Westgate Shopping Centre 7034 114029 6.17East Rand Mall 9076 147378 6.16The Zone @ Rosebank 5643 100035 5.64Eastgate Shopping Centre 8440 154396 5.47Key West Shopping Centre 5671 116502 4.87China Mall West 5465 116370 4.7Sandton City 1309 29601 4.42Melrose Arch 3203 82871 3.87Trade Route Mall 4385 131962 3.32Jabulani Mall 11382 353851 3.22Maponya Mall 5556 250246 2.22Dobsonville Shopping Centre 4107 227348 1.81
PREDICTIVELOGISTIC REGRESSION
• Age
• Job
• Contact method
• Furniture / Clothing accounts
• Previous marketing campaigns
• Current campaign
PREDICTOR VARIABLES
WHAT WE KNOW ABOUT THE LEADS
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
Model ‘Score’
Succ
ess
Rat
e (R
ealit
y)MODEL PREDICTION PERFORMANCE
Identify data that is relevant to
the business goal
Integrate and enrich the data
into an analytical data set
Run predictive algorithms
to find the model
Test the model to make
sure it will work
Use the model
in applications
Measure the effectiveness of
the model in the real world
Predictiveanalytics
Understand data
Monitor
Deploy
Evaluate
Prepare data
Model
FEEDBACK CYCLEIn Theory
Understand data
Monitor
Deploy
Evaluate
Prepare data
Model
FEEDBACK CYCLEIn Practice
Survey
Model
Score (Evaluate)
THANK YOU.