Course — Classification, Modern Regression and Multivariate
Exploration Using R
Here, I comment on motivations for this course.
Data Analysis Demands in 2013
"If you are looking for a career where your services will be in high
demand, you should find something where you provide a scarce,
complementary service to something that is getting ubiquitous and
cheap. So what's getting ubiquitous and cheap? Data. And what is
complementary to data? Analysis. So my recommendation is to take lots
of courses about how to manipulate and analyze data: databases,
machine learning, econometrics, statistics, visualization, and so
on." [Hal Varian, Chief Economist at Google]
Read the whole interview
Varian speaks - Youtube video
A new Twist on Data Deployment and Analysis
Data mining gives a new twist on data deployment and analysis
methodologies that have been developed over the past century or
more. A good overview is the online article by Mike Loukides at O'REILLY radar:
What is data science?
"The future belongs to the companies and people that turn data into products."
Technological and methodological changes and advances have included:
Huge increases in computational power and in computer storage
A synergy between theoretical and algorithmic advances, advances
in software and in computational power
Integration of what were formerly stand-alone abilities into single
software systems with a single interface and command language.
(The R system is a prime example; see below)
New types of data, and new opportunites for collecting data, arising
from advances in instrumentation, from the internet and from widespread
deployment of databases. Chapter 5 of Ayres (2007), entitled "Why now?",
has interesting commentary on the impact of such advances.
Classification is a major pre-occupation of data mining, with a more
limited focus on regression with a continuous outcome variable.
The aim is typically prediction rather than the more challenging
task of interpretation of model parameters.
The R System
Students intending to take this course will it useful to come with
some initial familiarity with the open-source
R system for scientific and
statistical computing and for graphics. The R system
is available without charge for
downloading from the internet.
It is a marvelous example of what can be achieved when highly
skilled specialists co-operate internationally, using the internet for
communication and co-ordination.
Background Reading
Ian Ayres 2007, Super Crunchers. Why Thinking-By-Numbers is the New
Way to be Smart. Bantam.
[This places data mining in a wider context of data-based
decision-making in business, government and consumer affairs. While
popular in style and short on analysis detail, it offers a useful
overview of ways in which applications of data mining and related
analytical techniques are developing and changing, in part because of
the new opportunities and challenges of the internet.]
Thomas H. Davenport and Jeanne G. Harris 2007, Competing on
Analytics: The New Science of Winning. Harvard Business School
Press. [Analytics is a buzzword for the application of data
mining type approaches in commerce. Davenport and Thomas give a
useful overview of issues for the deployment of analytical techniques
within organizations - benefits and traps, choice of amenable tasks,
the role of management, skill base issues, etc.]
Nate Silver: The Signal and the Noise. The Art and Science of
Prediction [Nate Silver called the outcome of last two US
presidential campaigns to within a whisker. Important points about
practical data analysis issues are well made. He covers the
financial crisis that started in 2010, prediction of election
results, picking likely sports stars, weather and climate
forecasting, earthquake prediction, economic forecasting, prediction
of epidemics, amd much more. He documents the recent record of
successes and failures of prediction in these areas.]
John Maindonald and John Braun 2010, Data Analysis and Graphics Using R - An Example-Based Approach, 3nd edn Cambridge University Press.
[Of greatest relevance to the course are Chapter 2
on Styles of Data Analysis, Chapters 5 & 6 (through to 6.3) on Linear Models,
Chapter 8 (through to 8.3) on logistic regression, Chapter 11 on Tree-based
Methods, and Chapter 12 (through to 12.2) on Multivariate Data Exploration &
Discrimination.]