Analytics Education
in the Era of Big Data
    Gregory Piatetsky
       KDnuggets


         © KDnuggets 2013   1
Outline
• Analytics, Data Mining, Data Science - What
  do we call it?
• Big Data Trends
• Jobs and Skills
• Analytics Education overview




                    © KDnuggets 2013            2
What do we call it?

   Buzzwords and Trends
    Useful for Marketing
“Analytics” has staying power

           (c) KDnuggets 2013   3
What do we call it?
• Statistics             Same Core Idea:
• Data mining            Finding Useful
• Knowledge Discovery in
  Data (KDD)             Patterns in Data
•   Predictive Analytics
•   Business Analytics
•   Data Science                       Different
•   Data Analytics                     Emphasis
•   …?

                           © KDnuggets 2013        4
Pre-history (1900-2000): Statistics




  statistics is the biggest term in 20th century,
  Analytics is used increasingly thru 20th century
  data mining appears in late 1990s
          From Google Ngram viewer – English language books
          Search case sensitive – used most popular spelling.

                                 © KDnuggets 2012               5
20th Century
     Analytics vs Data Mining


                                                     data mining




                                                     Data Mining
                                     Analytics

                                      analytics

??
         Google N-grams search is case sensitive;
         Note: “data mining” > “Data Mining” usage
         While “analytics” < “Analytics”
                  © KDnuggets 2012                                 6
Recent History: 1980-2008
                                                                    data mining




                         Analytics

                     analytics                                Knowledge Discovery

“analytics” has been used since 1900, but started to rise in 2005
“data mining” surges around 1995 (soon after first KDD conference) but slowly
declines after 2003 (TIA controversy, associated with Govt invasion of privacy).
“Knowledge Discovery” appears in 1989, rises in 1996, and plateaus in 2000
(Google N-grams, smoothing =1)
                                  © KDnuggets 2012                                 7
Google Trends:
After 2006, Analytics > Data Mining




             Global – all regions
               (c) KDnuggets 2012     8
>50% of “Analytics” searches are for
        “Google Analytics”




Google Analytics introduced,
Dec 2005
                               (c) KDnuggets 2012   9
Google Trends observations
         (as of Jan 2013) Decline in analytics in
                                                  2012?
data mining: 16   analytics -google: 54




        Competing on Analytics book,
        Tom Davenport, Apr 2007                 Vacation drops
                           (c) KDnuggets 2013
Global View: searches for
data mining, analytics -google




        Google Trends

               (c) KDnuggets 2013   11
Google Trends: USA, 2012
  For “analytics – google –web –adsense”




                  © KDnuggets 2013
                                           12
Google Trends:
USA, Regional Interest




        © KDnuggets 2013   13
Google Trends:
USA, Analytics-related terms




           © KDnuggets 2013    14
Analytics:
Business > Data> Predictive > Text




       Google Insights, Jan 2007- Sep 2012, Global
                     (c) KDnuggets 2012              15
Big Data > Data Mining >
* Analytics > Data Science




                                      “Big Data” Surge




   Google Trends search, Jan 2007- Dec 2012, USA

                 (c) KDnuggets 2013                      16
Big Data Trends




    (c) KDnuggets 2013   17
3 Vs of Big Data
• Volume
  – Gigabytes to Terabytes to Petabytes …
• Velocity
  – online streaming
• Variety
  – numbers, text, links, images, audio, video, …




                       (c) KDnuggets 2013           18
Volume + Velocity => No consistency
• CAP Theorem (Eric Brewer, 2000)
  For highly scalable distributed systems, you can only
    have two of following:
  – 1) consistency,
  – 2) high availability, and
  – 3) (network) partition tolerance (network failure tolerance)
   http://www.julianbrowne.com/article/viewer/brewers-cap-
     theorem

  Implication: Big data solutions must stop worrying
    about consistency if they want high availability

                           (c) KDnuggets 2013                      19
Big Data
• 2nd Industrial Revolution

• Do old activities better

• Create new activities/businesses




                     (c) KDnuggets 2013   20
Doing Old Things Better
“Classical” Analytics Application areas
  – Churn prediction
  – Direct marketing/Customer modeling
  – Recommendations
  – Fraud detection
  – Security/Intelligence
  –…
• Competition will level companies

                     (c) KDnuggets 2013   21
Limit to Predicting Human Behavior?
• There is randomness in human behavior and
  once we find first-level effects, there are
  diminishing returns in prediction on individual
  level
• Many examples: Netflix Prize, Customer
  modeling…

Gregory Piatetsky-Shapiro, Big Data Hype and Reality,
  Harvard Business Review blog, Oct 2012

                       (c) KDnuggets 2013               22
Netflix Prize Progress
The most advanced algorithms were only a few
  percentages better than basic algorithms




                   © KDnuggets 2013            23
Many Customer Modeling Tasks have
            Similar Lift
                    Actual lift(T)    Est. lift(T)                Lift(T) ~ T -0.5 = sqrt (1/T)
       14
       12
                                                                  See G. Piatetsky-Shapiro,
       10                                                         B. Masand, Estimating
        8                                                         Campaign Benefits and
Lift




        6                                                         Modeling Lift,
        4                                                         Proceedings of KDD-99
        2
                                                                  Conference
        0
            0   5       10           15           20         25
                          100*T%




                                          © KDnuggets 2013                                 24
Big Data Bubble?                  Gartner VP says Big Data
                                  is Falling into the Trough
                                  of Disillusionment, Jan
                                  2013


   Big Data




   Gartner Hype Cycle

                                                          25
               © 2013 KDnuggets
Big Data Enables New Things !
– Google – first big success of big data
– Social networks (Facebook, Twitter, LinkedIn, …)
  success depends on network size, i.e. big data

– Location analytics
– Health-care
   • Personalized medicine
– Semantics and AI ?
   • Imagine IBM Watson, Siri in 2020 ?
– Beware of Loss of privacy

                       (c) KDnuggets 2012            26
Largest Dataset Analyzed?
                                                    Big Data Miners – elite
                                                    group

                                                    2012 median dataset size
                                                    ~20-40 GB,
                                                    vs 10-20 GB in 2011.




www.KDnuggets.com/polls/2012/largest-dataset-analyzed-data-mined.html
                             © KDnuggets 2012                             27
Digital Universe in 2020

                                            40,000 exabytes by 2020

                                            1 Exabyte = 1018 bytes
                                            = 1,000,000 TB




   Source : IDC Study The Digital Universe in 2020
                  © KDnuggets 2013                               28
Where in the World is Big Data?
                                Most of Big Data
                                is outside the US




             © KDnuggets 2012                       29
Opportunities for Big Data
• Social networks
• Surveillance video
• Embedded and medical devices
  – M2M data, logs
• Entertainment and social media
• Consumer images – image recognition,
  labeling, …


                     © KDnuggets 2012    30
Where did you apply Analytics/Data Mining?

                                                  Avg. Number of Industries 2.6

                                                  Most Popular:
                                                  - CRM/Consumer analytics
                                                  - Health care/ HR,
                                                  - Retail
                                                  - Banking
                                                  - Education

                                                  Highest growth in:
                                                  1. Advertising, 89.0%
                                                  2. Search / Web content
                                                      mining, 55.1%
                                                  3. Retail, 40.6%
                                                  4. Other, 36.9%
                                                  5. Manufacturing, 35.7%

www.KDnuggets.com/polls/2012/where-applied-analytics-data-mining.html
                               © KDnuggets 2013                              31
Untapped Big Data Gap




 Big limitation is lack of Analytic Talent
              © KDnuggets 2013               32
JOBS AND SKILLS



    (c) KDnuggets 2011   33
Shortage of Skills
• McKinsey: shortage by 2018 in the US of
  – 140-190,000 people with deep analytical skills

  – 1.5 M managers/analysts with the know-how to
    use the analysis of big data to make effective
    decisions.

  Source:
    www.mckinsey.com/mgi/publications/big_data/

                      (c) KDnuggets 2012             34
Indeed.com fastest growing jobs
Top 10 skills:
• HTML5                             Hadoop
• MongoDB
• iOS
• Android                                    MongoDB
• Mobile app
• Puppet
• Hadoop
• jQuery
• PaaS
• Social Media

                 © KDnuggets 2012               35
“Big Data” grows faster than MongoDB

                                  Big Data

                                  Hadoop




                                  MongoDB




               © KDnuggets 2012        36
Data Mining >> Hadoop




        (c) KDnuggets 2013   37
Demand for Data Scientists surging

                                                           “Data Scientist”

                                                           Fastest growing term on
                                                           www.kdnuggets.com/jobs

                                                           1% of jobs in 2010

                                                           4% of jobs in 2011

                                                           19% of jobs in 2012



 Data Scientist – sexiest job of the 21st Century (???)
 say Thomas H. Davenport and D.J. Patil, (HBR, Oct 2012)

                               © KDnuggets 2013                                  38
What is a Data Scientist
My definition: A combination of MBA, a Statistician, and a Hacker




                                                         Drew Conway:
                                                         http://www.drewconway.co
                                                         m/zia/?p=2378


                                  © KDnuggets 2013                            39
Rebranding from
“Data Mining” to “Big Data”


                  Data Mining




                                                  Big Data

                                                             Data Scientist


  “Data mining” jobs are much more common, but
  “Big Data” jobs are surging much faster than “Data Scientist”

                    (c) KDnuggets 2011                                40
LinkedIn Analytics/Data Mining Skills
                                    “Ground” analytics
                                    skills most common

                                    “Cloud” analytics
                                    skills grow
                                    fastest

                                    Text Analytics skills
                                    less common
                                    Sentiment Analysis
                                    – fastest growing




               (c) KDnuggets 2012                       41
Analytics Education




      © KDnuggets 2013   42
Analytics Education: USA/Canada
US: Northeast | US: South | US: Midwest | US: West | Canada
     Northeast
•    Connecticut: Central Connecticut State University (CCSU), exploring cutting-edge data mining techniques
     and Applications. New Britain, CT.
•     U. Conn. MS in Business Analytics and Project Management, designed to meet the growing demand for
     professionals who can harness advanced business analytics and project management skills. Hartford, CT.
•    Maryland: U. of Maryland MS in Business for Marketing Analytics, will help you learn how to harness and
     process massive amounts of data to help design products, predict the effects of marketing campaigns, and
     better understand your customers. Fall 2013. College Park, MD.
•    Massachusetts: Bentley Master of Science in Marketing Analytics, teaches students how to become more
     engaged with consumers, how to design and deliver robust statistical analysis, and how to effectively
     communicate the resulting insights. Waltham, MA.
•     Harvard Masters of Science in Computational Science and Engineering, including a major focus on
     machine learning and analyzing and visualizing very large data sets. Cambridge, MA.
•    New Jersey: Rutgers Master of Business and Science (MBS) in Analytics, prepares students for data-driven
     decision making; brings together fields of data management, statistics, machine learning and
     computation. New Brunswick, NJ.
•    Stevens Institute of Technology Master of Science - Business Intelligence & Analytics, Hoboken, NJ.
•    New York: Columbia MS in Computer Science, concentration in Machine Learning, New York, NY.
•     NYU Master of Science in Business Analytics, starting May 2013, New York, NY.
•     NYU MBA with specialization in Business Analytics, New York, NY.
•    ….
Full list at http://www.kdnuggets.com/education/usa-canada.html


                                              © KDnuggets 2013                                             43
www.kdnuggets.com/education/usa-canada.html
Analytics Education: Online
•   Big Data University, offering online classes on Hadoop and DB2.
•   Caltech Learning from Data course, free, broadcast online Apr-May 2012.
•   CMU Open Learning Initiative, including on-line courses in statistics, math & logic.
•   Coursera, offering online classes from Stanford and other top universities.
    Check especially Statistics, Data Analysis, and Scientific Computing courses.
•   Data Mining Tools Tutorials, covering Data Mining, Probability, Weka, R, and numerous commercial data mining tools.
•   EMC Data Science and Big Data Analytics open course.
•   LearnAnalytics India, delivering SAS and Advanced Analytics trainings online and offline.
•   Northwestern University Online Master of Science in Predictive Analytics, skills for leadership in a growing
    Field.
•   Oxford Advanced Diploma in Data and Systems Analysis, a one-year online course.
•    Stanford Center for Professional Education, offers certificate programs for managers and professionals in Data
    Mining and Applications and many related areas.
•   Statistics.com, offering on-line short courses in statistics and data mining.
•   UCI: U. of California, Irvine Extension, Predictive Analytics Certificate Program, a comprehensive online
    program.
•   UC San Diego Data Mining Courses, part of Data Analysis study area.
•   Udacity, online university founded by David Evans and Sebastian Thrun.
•   Video lectures from conferences, workshops and the scientific lectures in the areas of machine learning, data and text
    mining, and semantic web.


Full list at http://www.kdnuggets.com/education/online.html




                                                     © KDnuggets 2013                                                        44
Analytics Certificates
•   Business Analytics Certificates from BeyeUniversity, sponsored by Sybase.
•   Business Analytics Certificate from Statistics.com, become your company's expert on forecasting, customer segmentation, consumer
    behavior, and risk analysis.
•   Central Michigan University Graduate certificate in Data Mining, with SAS. Mount Pleasant, MI, USA
•   Data Mining Certificate from Statistics.com, master the secrets of teasing powerful information from large data sets to predict customer
    behavior, identify likely high value customers, visualize high dimensional data, and convert text to minable data.
•   EMC Data Science and Big Data Analytics open curriculum-based Education and Certification.
•    Indiana U. Kelley School of Business Business Analytics Certificate Program, Bloomington, IN, USA.
•   INFORMS Analytics Certification, coming April 2013.
•   NJIT (New Jersey Institute of Technology) The Graduate Certificate in Data Mining, online and in class.
•   Nova Southeastern University Graduate Certificate In Business Intelligence / Analytics, Fort Lauderdale, FL.
•   Stanford Center for Professional Education, offers certificate programs for managers and professionals in Data Mining and Applications and
    Quantitative Methods in Finance and Risk Management.
•   Statistics.com Institute for Statistics Education, offers certificates in Data Analytics, Biostatistics, Social Science, and Using R.
•   UCI: U. of California, Irvine Extension, Predictive Analytics Certificate Program, a comprehensive online program.
•   UCSD Data Mining certificate program, San Diego, CA.
•    University of Delaware Certificate in Analytics: Optimizing Big Data, Newark, DE, USA.
•   U. of Washington Certificate in Data Science, Seattle, WA, USA.
•   SAS Certificate Programs


Full list http://www.kdnuggets.com/education/analytics-data-mining-certificates.html



                                                            © KDnuggets 2013                                                                 45
Analytics Education by Doing
• Competitions – learn by doing
  – Kaggle and more
• Kaggle beginner competitions
• Kaggle in class : free to Instructors from any
  course dealing with data analysis




                      © KDnuggets 2013             46
Online Education – Free Courses
Data Science pseudo degree from
• Lower-Division Courses
   –   Data Science 101 – Statistics One
   –   Data Science 102 – Computing for Data Analysis (R)
   –   Data Science 103 – Data Analysis
   –   Data Science 104 – Introduction to Data Science
• Upper-Division Courses
   – Data Science 201 – Machine Learning I
   – Data Science 202 – Machine Learning II
   – Data Science 203 – Neural Networks for Machine Learning
• Graduate Courses
   – Data Science 301 – Learning from Data (Caltech course CS101)
   – Data Science 302 – Machine Learning III (MIT course 6.867)


                                © KDnuggets 2013                    47
Analytics Education:
 Practical Knowledge > Degree Prestige
Paco Nathan - School prestige matters some to hiring managers. Several top schools
   are known to have excellent programs and track records: Stanford, CMU, U
   Washington, UC Berkeley, Harvard, Johns Hopkins, etc. However, keep in mind that
   that list is not entirely representative. For example, there are nearly a dozen
   relevant programs at Stanford, which has produced Google, Yahoo, etc., while the
   list only mentions one program.

More to the point, about half of my peers in this field have backgrounds in Physics or
  physical science/physical engineering -- and I tend to hire from those programs
  more so than from CS programs, because the grad students tend to have both the
  math/stats depth plus practice with real-world frameworks like R, Matlab, etc.
  Having a really solid background in applying statistics at scale, some knack for data
  visualization, plus good programming chops -- those skills will trump a PhD in
  Machine Learning.

http://www.linkedin.com/groups/Graduate-Programs-in-Big-Data-
    2013423.S.203868026?view=&srchtype=discussedNews&gid=2013423&item=203
    868026&type=member&trk=eml-anet_dig-b_pd-ttl-cn&ut=2Xc36Ur0k0zBA1


                                     © KDnuggets 2013                                 48
Analytics Education Boom




2012 was a peak year for starting Analytics Degree Programs in the US


                           © KDnuggets 2013                             49
Questions?
Analytics Education Overview
www.kdnuggets.com/education/

Subscribe to KDnuggets News email at
  www.KDnuggets.com/subscribe.html

• Follow @kdnuggets on Twitter

• Email to editor1@kdnuggets.com
                  © KDnuggets 2013     50

Analytics Education in the era of Big Data

  • 1.
    Analytics Education in theEra of Big Data Gregory Piatetsky KDnuggets © KDnuggets 2013 1
  • 2.
    Outline • Analytics, DataMining, Data Science - What do we call it? • Big Data Trends • Jobs and Skills • Analytics Education overview © KDnuggets 2013 2
  • 3.
    What do wecall it? Buzzwords and Trends Useful for Marketing “Analytics” has staying power (c) KDnuggets 2013 3
  • 4.
    What do wecall it? • Statistics Same Core Idea: • Data mining Finding Useful • Knowledge Discovery in Data (KDD) Patterns in Data • Predictive Analytics • Business Analytics • Data Science Different • Data Analytics Emphasis • …? © KDnuggets 2013 4
  • 5.
    Pre-history (1900-2000): Statistics statistics is the biggest term in 20th century, Analytics is used increasingly thru 20th century data mining appears in late 1990s From Google Ngram viewer – English language books Search case sensitive – used most popular spelling. © KDnuggets 2012 5
  • 6.
    20th Century Analytics vs Data Mining data mining Data Mining Analytics analytics ?? Google N-grams search is case sensitive; Note: “data mining” > “Data Mining” usage While “analytics” < “Analytics” © KDnuggets 2012 6
  • 7.
    Recent History: 1980-2008 data mining Analytics analytics Knowledge Discovery “analytics” has been used since 1900, but started to rise in 2005 “data mining” surges around 1995 (soon after first KDD conference) but slowly declines after 2003 (TIA controversy, associated with Govt invasion of privacy). “Knowledge Discovery” appears in 1989, rises in 1996, and plateaus in 2000 (Google N-grams, smoothing =1) © KDnuggets 2012 7
  • 8.
    Google Trends: After 2006,Analytics > Data Mining Global – all regions (c) KDnuggets 2012 8
  • 9.
    >50% of “Analytics”searches are for “Google Analytics” Google Analytics introduced, Dec 2005 (c) KDnuggets 2012 9
  • 10.
    Google Trends observations (as of Jan 2013) Decline in analytics in 2012? data mining: 16 analytics -google: 54 Competing on Analytics book, Tom Davenport, Apr 2007 Vacation drops (c) KDnuggets 2013
  • 11.
    Global View: searchesfor data mining, analytics -google Google Trends (c) KDnuggets 2013 11
  • 12.
    Google Trends: USA,2012 For “analytics – google –web –adsense” © KDnuggets 2013 12
  • 13.
    Google Trends: USA, RegionalInterest © KDnuggets 2013 13
  • 14.
    Google Trends: USA, Analytics-relatedterms © KDnuggets 2013 14
  • 15.
    Analytics: Business > Data>Predictive > Text Google Insights, Jan 2007- Sep 2012, Global (c) KDnuggets 2012 15
  • 16.
    Big Data >Data Mining > * Analytics > Data Science “Big Data” Surge Google Trends search, Jan 2007- Dec 2012, USA (c) KDnuggets 2013 16
  • 17.
    Big Data Trends (c) KDnuggets 2013 17
  • 18.
    3 Vs ofBig Data • Volume – Gigabytes to Terabytes to Petabytes … • Velocity – online streaming • Variety – numbers, text, links, images, audio, video, … (c) KDnuggets 2013 18
  • 19.
    Volume + Velocity=> No consistency • CAP Theorem (Eric Brewer, 2000) For highly scalable distributed systems, you can only have two of following: – 1) consistency, – 2) high availability, and – 3) (network) partition tolerance (network failure tolerance) http://www.julianbrowne.com/article/viewer/brewers-cap- theorem Implication: Big data solutions must stop worrying about consistency if they want high availability (c) KDnuggets 2013 19
  • 20.
    Big Data • 2ndIndustrial Revolution • Do old activities better • Create new activities/businesses (c) KDnuggets 2013 20
  • 21.
    Doing Old ThingsBetter “Classical” Analytics Application areas – Churn prediction – Direct marketing/Customer modeling – Recommendations – Fraud detection – Security/Intelligence –… • Competition will level companies (c) KDnuggets 2013 21
  • 22.
    Limit to PredictingHuman Behavior? • There is randomness in human behavior and once we find first-level effects, there are diminishing returns in prediction on individual level • Many examples: Netflix Prize, Customer modeling… Gregory Piatetsky-Shapiro, Big Data Hype and Reality, Harvard Business Review blog, Oct 2012 (c) KDnuggets 2013 22
  • 23.
    Netflix Prize Progress Themost advanced algorithms were only a few percentages better than basic algorithms © KDnuggets 2013 23
  • 24.
    Many Customer ModelingTasks have Similar Lift Actual lift(T) Est. lift(T) Lift(T) ~ T -0.5 = sqrt (1/T) 14 12 See G. Piatetsky-Shapiro, 10 B. Masand, Estimating 8 Campaign Benefits and Lift 6 Modeling Lift, 4 Proceedings of KDD-99 2 Conference 0 0 5 10 15 20 25 100*T% © KDnuggets 2013 24
  • 25.
    Big Data Bubble? Gartner VP says Big Data is Falling into the Trough of Disillusionment, Jan 2013 Big Data Gartner Hype Cycle 25 © 2013 KDnuggets
  • 26.
    Big Data EnablesNew Things ! – Google – first big success of big data – Social networks (Facebook, Twitter, LinkedIn, …) success depends on network size, i.e. big data – Location analytics – Health-care • Personalized medicine – Semantics and AI ? • Imagine IBM Watson, Siri in 2020 ? – Beware of Loss of privacy (c) KDnuggets 2012 26
  • 27.
    Largest Dataset Analyzed? Big Data Miners – elite group 2012 median dataset size ~20-40 GB, vs 10-20 GB in 2011. www.KDnuggets.com/polls/2012/largest-dataset-analyzed-data-mined.html © KDnuggets 2012 27
  • 28.
    Digital Universe in2020 40,000 exabytes by 2020 1 Exabyte = 1018 bytes = 1,000,000 TB Source : IDC Study The Digital Universe in 2020 © KDnuggets 2013 28
  • 29.
    Where in theWorld is Big Data? Most of Big Data is outside the US © KDnuggets 2012 29
  • 30.
    Opportunities for BigData • Social networks • Surveillance video • Embedded and medical devices – M2M data, logs • Entertainment and social media • Consumer images – image recognition, labeling, … © KDnuggets 2012 30
  • 31.
    Where did youapply Analytics/Data Mining? Avg. Number of Industries 2.6 Most Popular: - CRM/Consumer analytics - Health care/ HR, - Retail - Banking - Education Highest growth in: 1. Advertising, 89.0% 2. Search / Web content mining, 55.1% 3. Retail, 40.6% 4. Other, 36.9% 5. Manufacturing, 35.7% www.KDnuggets.com/polls/2012/where-applied-analytics-data-mining.html © KDnuggets 2013 31
  • 32.
    Untapped Big DataGap Big limitation is lack of Analytic Talent © KDnuggets 2013 32
  • 33.
    JOBS AND SKILLS (c) KDnuggets 2011 33
  • 34.
    Shortage of Skills •McKinsey: shortage by 2018 in the US of – 140-190,000 people with deep analytical skills – 1.5 M managers/analysts with the know-how to use the analysis of big data to make effective decisions. Source: www.mckinsey.com/mgi/publications/big_data/ (c) KDnuggets 2012 34
  • 35.
    Indeed.com fastest growingjobs Top 10 skills: • HTML5 Hadoop • MongoDB • iOS • Android MongoDB • Mobile app • Puppet • Hadoop • jQuery • PaaS • Social Media © KDnuggets 2012 35
  • 36.
    “Big Data” growsfaster than MongoDB Big Data Hadoop MongoDB © KDnuggets 2012 36
  • 37.
    Data Mining >>Hadoop (c) KDnuggets 2013 37
  • 38.
    Demand for DataScientists surging “Data Scientist” Fastest growing term on www.kdnuggets.com/jobs 1% of jobs in 2010 4% of jobs in 2011 19% of jobs in 2012 Data Scientist – sexiest job of the 21st Century (???) say Thomas H. Davenport and D.J. Patil, (HBR, Oct 2012) © KDnuggets 2013 38
  • 39.
    What is aData Scientist My definition: A combination of MBA, a Statistician, and a Hacker Drew Conway: http://www.drewconway.co m/zia/?p=2378 © KDnuggets 2013 39
  • 40.
    Rebranding from “Data Mining”to “Big Data” Data Mining Big Data Data Scientist “Data mining” jobs are much more common, but “Big Data” jobs are surging much faster than “Data Scientist” (c) KDnuggets 2011 40
  • 41.
    LinkedIn Analytics/Data MiningSkills “Ground” analytics skills most common “Cloud” analytics skills grow fastest Text Analytics skills less common Sentiment Analysis – fastest growing (c) KDnuggets 2012 41
  • 42.
    Analytics Education © KDnuggets 2013 42
  • 43.
    Analytics Education: USA/Canada US:Northeast | US: South | US: Midwest | US: West | Canada Northeast • Connecticut: Central Connecticut State University (CCSU), exploring cutting-edge data mining techniques and Applications. New Britain, CT. • U. Conn. MS in Business Analytics and Project Management, designed to meet the growing demand for professionals who can harness advanced business analytics and project management skills. Hartford, CT. • Maryland: U. of Maryland MS in Business for Marketing Analytics, will help you learn how to harness and process massive amounts of data to help design products, predict the effects of marketing campaigns, and better understand your customers. Fall 2013. College Park, MD. • Massachusetts: Bentley Master of Science in Marketing Analytics, teaches students how to become more engaged with consumers, how to design and deliver robust statistical analysis, and how to effectively communicate the resulting insights. Waltham, MA. • Harvard Masters of Science in Computational Science and Engineering, including a major focus on machine learning and analyzing and visualizing very large data sets. Cambridge, MA. • New Jersey: Rutgers Master of Business and Science (MBS) in Analytics, prepares students for data-driven decision making; brings together fields of data management, statistics, machine learning and computation. New Brunswick, NJ. • Stevens Institute of Technology Master of Science - Business Intelligence & Analytics, Hoboken, NJ. • New York: Columbia MS in Computer Science, concentration in Machine Learning, New York, NY. • NYU Master of Science in Business Analytics, starting May 2013, New York, NY. • NYU MBA with specialization in Business Analytics, New York, NY. • …. Full list at http://www.kdnuggets.com/education/usa-canada.html © KDnuggets 2013 43 www.kdnuggets.com/education/usa-canada.html
  • 44.
    Analytics Education: Online • Big Data University, offering online classes on Hadoop and DB2. • Caltech Learning from Data course, free, broadcast online Apr-May 2012. • CMU Open Learning Initiative, including on-line courses in statistics, math & logic. • Coursera, offering online classes from Stanford and other top universities. Check especially Statistics, Data Analysis, and Scientific Computing courses. • Data Mining Tools Tutorials, covering Data Mining, Probability, Weka, R, and numerous commercial data mining tools. • EMC Data Science and Big Data Analytics open course. • LearnAnalytics India, delivering SAS and Advanced Analytics trainings online and offline. • Northwestern University Online Master of Science in Predictive Analytics, skills for leadership in a growing Field. • Oxford Advanced Diploma in Data and Systems Analysis, a one-year online course. • Stanford Center for Professional Education, offers certificate programs for managers and professionals in Data Mining and Applications and many related areas. • Statistics.com, offering on-line short courses in statistics and data mining. • UCI: U. of California, Irvine Extension, Predictive Analytics Certificate Program, a comprehensive online program. • UC San Diego Data Mining Courses, part of Data Analysis study area. • Udacity, online university founded by David Evans and Sebastian Thrun. • Video lectures from conferences, workshops and the scientific lectures in the areas of machine learning, data and text mining, and semantic web. Full list at http://www.kdnuggets.com/education/online.html © KDnuggets 2013 44
  • 45.
    Analytics Certificates • Business Analytics Certificates from BeyeUniversity, sponsored by Sybase. • Business Analytics Certificate from Statistics.com, become your company's expert on forecasting, customer segmentation, consumer behavior, and risk analysis. • Central Michigan University Graduate certificate in Data Mining, with SAS. Mount Pleasant, MI, USA • Data Mining Certificate from Statistics.com, master the secrets of teasing powerful information from large data sets to predict customer behavior, identify likely high value customers, visualize high dimensional data, and convert text to minable data. • EMC Data Science and Big Data Analytics open curriculum-based Education and Certification. • Indiana U. Kelley School of Business Business Analytics Certificate Program, Bloomington, IN, USA. • INFORMS Analytics Certification, coming April 2013. • NJIT (New Jersey Institute of Technology) The Graduate Certificate in Data Mining, online and in class. • Nova Southeastern University Graduate Certificate In Business Intelligence / Analytics, Fort Lauderdale, FL. • Stanford Center for Professional Education, offers certificate programs for managers and professionals in Data Mining and Applications and Quantitative Methods in Finance and Risk Management. • Statistics.com Institute for Statistics Education, offers certificates in Data Analytics, Biostatistics, Social Science, and Using R. • UCI: U. of California, Irvine Extension, Predictive Analytics Certificate Program, a comprehensive online program. • UCSD Data Mining certificate program, San Diego, CA. • University of Delaware Certificate in Analytics: Optimizing Big Data, Newark, DE, USA. • U. of Washington Certificate in Data Science, Seattle, WA, USA. • SAS Certificate Programs Full list http://www.kdnuggets.com/education/analytics-data-mining-certificates.html © KDnuggets 2013 45
  • 46.
    Analytics Education byDoing • Competitions – learn by doing – Kaggle and more • Kaggle beginner competitions • Kaggle in class : free to Instructors from any course dealing with data analysis © KDnuggets 2013 46
  • 47.
    Online Education –Free Courses Data Science pseudo degree from • Lower-Division Courses – Data Science 101 – Statistics One – Data Science 102 – Computing for Data Analysis (R) – Data Science 103 – Data Analysis – Data Science 104 – Introduction to Data Science • Upper-Division Courses – Data Science 201 – Machine Learning I – Data Science 202 – Machine Learning II – Data Science 203 – Neural Networks for Machine Learning • Graduate Courses – Data Science 301 – Learning from Data (Caltech course CS101) – Data Science 302 – Machine Learning III (MIT course 6.867) © KDnuggets 2013 47
  • 48.
    Analytics Education: PracticalKnowledge > Degree Prestige Paco Nathan - School prestige matters some to hiring managers. Several top schools are known to have excellent programs and track records: Stanford, CMU, U Washington, UC Berkeley, Harvard, Johns Hopkins, etc. However, keep in mind that that list is not entirely representative. For example, there are nearly a dozen relevant programs at Stanford, which has produced Google, Yahoo, etc., while the list only mentions one program. More to the point, about half of my peers in this field have backgrounds in Physics or physical science/physical engineering -- and I tend to hire from those programs more so than from CS programs, because the grad students tend to have both the math/stats depth plus practice with real-world frameworks like R, Matlab, etc. Having a really solid background in applying statistics at scale, some knack for data visualization, plus good programming chops -- those skills will trump a PhD in Machine Learning. http://www.linkedin.com/groups/Graduate-Programs-in-Big-Data- 2013423.S.203868026?view=&srchtype=discussedNews&gid=2013423&item=203 868026&type=member&trk=eml-anet_dig-b_pd-ttl-cn&ut=2Xc36Ur0k0zBA1 © KDnuggets 2013 48
  • 49.
    Analytics Education Boom 2012was a peak year for starting Analytics Degree Programs in the US © KDnuggets 2013 49
  • 50.
    Questions? Analytics Education Overview www.kdnuggets.com/education/ Subscribeto KDnuggets News email at www.KDnuggets.com/subscribe.html • Follow @kdnuggets on Twitter • Email to [email protected] © KDnuggets 2013 50

Editor's Notes

  • #19 Boris Evelson, Forrester also adds 4th V – Variability (meaning not constant)
  • #23 Churn: bestalgorithms for predicting churn have lift of 5-7 – 5-7 times better than random. Behavioral advertising: 2-3% CTR – 10 times better than random