100% found this document useful (1 vote)
206 views

Data Science: Where Does A Physics Degree Fit?

The presentation discusses the relevance of scientific methods to digital data. For more information, visit <a href="http://www.datalicous.com">Datalicious</a>
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
206 views

Data Science: Where Does A Physics Degree Fit?

The presentation discusses the relevance of scientific methods to digital data. For more information, visit <a href="http://www.datalicous.com">Datalicious</a>
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 31

>

Data Science <


Where does a Physics degree t?

> Life as a Data Scien/st


What is it? Data Science Examples Why you should think about it? About Datalicious (my current company)

August 2011

Datalicious Pty Ltd

101011010010010010101111010010010101010100001011111001010101 010100101011001100010100101001101101001101001010100111001010 010010101001001010010100100101001111101010100101001001001010

> What is it?


August 2011 Datalicious Pty Ltd 3

> Mixing the old to produce new


Business Performance Marke/ng / Adver/sing

IT / Coding

Sta/s/cs

Data Science
August 2011 Datalicious Pty Ltd 4

> By Deni/on
Data: Facts and sta)s)cs collected together for reference or analysis

Science: The systema)c study of the structure and behaviour of the physical and natural world through observa)on and experiment Hmmm, so what?
August 2011 Datalicious Pty Ltd 5

> Sounds like nothing new


Data, Science, ObservaPon, Hypothesis, Experiment, Analysis, PredicPon are all nothing new. BUT, the digital age has created new opportuniPes where scienPc methods can be applied to massive, real world digital data sets.
August 2011 Datalicious Pty Ltd 6

> And Data is Exploding


Every 2 Days We Create As Much Informa)on As We Did Up To 2003 Eric Schmidt. CEO, Google AdMob Seeing 2 Billion Ad Requests Per Day; Up 300 Percent Over Past Year TechCrunch The amount of digital informa)on created annually will grow by a factor of 44 from 2009 to 2020, as all major forms of media - voice, TV, radio, print - complete the journey from analog to digital
August 2011 Datalicious Pty Ltd 7

> EMC - "The Digital Universe Decade - Are You Ready?"


In 2009, amid the "Great Recession," the amount of digital informa)on grew 62% over 2008 to 800 billion gigabytes (0.8 Ze\abytes). But how much is that really? 707 trillion copies of the more than 2,000-page U.S. PaPent ProtecPon and Aordable Care Act signed into Law in March 2010. Stacked end to end, the documents would stretch from Earth to Pluto and back 16 Pmes or cover every inch of the United States in paper 3 feet deep
August 2011 Datalicious Pty Ltd

[ Mo/va/on ]

Data =
Datalicious Pty Ltd 9

101011010010010010101111010010010101010100001011111001010101 010100101011001100010100101001101101001101001010100111001010 010010101001001010010100100101001111101010100101001001001010

> Data Science Examples


August 2011 Datalicious Pty Ltd 10

> Search and the product lifecycle


Nokia N-Series

Apple iPhone
October 2010 Datalicious Pty Ltd 11

> Financial Predic/on

Datalicious Pty Ltd

12

> Trigger based Sales


User visits Site anonymously Iden/es themself (e.g. sale or registra/on) 2 years on the beach User visits Site again anonymously Looks at lots of widgets

Cookie Web Analy/cs Database Name: John Example Interest: Widgets History: -Last visit 2 years ago. -Purchased 10 blue widgets -High value band Loca/on: 2000, Sydney

Cookie

Hi John, long /me no talk, we have a special on widgets!

Business Intelligence Database

Datalicious Pty Ltd

13

> The science of Banner Ads


Control Group (10%) Normal Display Adver/sing (90%)

Impressions = 1 M Visitors = 1,000 Sales = 50 Cost = $1,000 $ per sale = $20 Without display

adverPsing, 1 in every 20,000 people will convert

22%
uplie

With display adverPsing 1 in every 16,363 people will convert

Impressions = 9 M Visitors = 10,000 Sales = 550 Cost = $9,000 $ per sale = $16

- Banner Ads do inuence sales despite what people think


- Even if you dont click on them
Datalicious Pty Ltd 14

[ Anity targe/ng in ac/on ]


Dierent types of visitors respond to dierent ads. By using category anity targePng, response rates are lieed signicantly across products.
Message Blackberry Bold 5GB Mobile Broadband Blackberry Storm 12 Month Caps Click-Through Rate By Category Anity Postpay Prepay Broadb. Business

+ -

+ + -

+ + +
15

Datalicious Pty Ltd

14/11/12

Datalicious Pty Ltd

16

101011010010010010101111010010010101010100001011111001010101 010100101011001100010100101001101101001101001010100111001010 010010101001001010010100100101001111101010100101001001001010

> Why should you do it?


August 2011 Datalicious Pty Ltd 17

> The start of a trend

August 2011

Datalicious Pty Ltd

18

> Big Data also

August 2011

Datalicious Pty Ltd

19

> Supply is way behind demand

August 2011

Datalicious Pty Ltd

20

IT needs more Scien/sts

August 2011

Datalicious Pty Ltd

21

> Dicult to believe?


Marketers typically just dont get it, but their bosses now know you can measure the ROI of digital Ads, so theyre screwed without the data scienPsts The business guys mostly (not all) see the value, but the Internet wasnt around when they were back in Harvard, so they cant do it alone

August 2011

Developers are predominantly good at following a spec. They rarely understand data, staPsPcs and how to go about sejng up and analysing an experiment Datalicious Pty Ltd 22

> You can work across all industries

August 2011

Datalicious Pty Ltd

23

> Work on wide ranging issues

August 2011

Datalicious Pty Ltd

24

> Work from Anywhere

August 2011

Datalicious Pty Ltd

25

101011010010010010101111010010010101010100001011111001010101 010100101011001100010100101001101101001101001010100111001010 010010101001001010010100100101001111101010100101001001001010

> About Datalicious


August 2011 Datalicious Pty Ltd 26

> Across major data categories


Campaign data TV, print, call center, search, web analyPcs, ad serving, etc Campaigns Customers

Customer data Direct mail, call center, web analyPcs, emails, surveys, etc

Compe/tors

Consumers

Consumer data Search, social media, trends, research, news, etc


Compe/tor data Search, social media, ad spend, news, oers, etc


August 2011 Datalicious Pty Ltd 27

> Dening data strategies


Con/nuous tes/ng and op/miza/on

AcquisiPon

Up-Sell

RetenPon

Advocacy

Analy/cs and metrics frameworks


August 2011 Datalicious Pty Ltd 28

> Guiding the customer journey


To transac/onal data To reten/on messages

From suspect to
Time

prospect

To customer
Time

From behavioural data


August 2011 Datalicious Pty Ltd

From awareness messages


29

> Summary
Do you:
1. Know how to idenPfy trends in numbers and to graph data 2. Know how to write reports and validate experimental predicPons 3. Understand some business thinking, i.e. cost of sales, maximising return, etc 4. Understand the principles of wriPng code

If yes, then Data Science may be for you.


August 2011 Datalicious Pty Ltd 30

[email protected] blog.datalicious.com twiner.com/hamishogilvy twiner.com/datalicious


August 2011 Datalicious Pty Ltd 31

Contact us Learn

Follow

You might also like