Diapositiva 1

1
Data Mining and visualization (1) Alfredo
Vellido
2
Plan (for parts 1 2)

• A brief introduction to data visualization
• Visualization history
• Perception seeing with the brain
• Visual exploratory DM
• The good, the bad the ugly

3
Maps and cartograms of the 2004 US presidential
election results

• The map most press published after USA04
  presidential elections

Republicans Democrats
www-personal.umich.edu/mejn/election/
Michael Gastner, Cosma Shalizi, and Mark Newman
(University of Michigan)
4
Maps and cartograms of the 2004 US presidential
election results (2)

• which is not the same as a cartogram, corrected
  according to state population

5
Maps and cartograms of the 2004 US presidential
election results (3)

• but it happens that the US election results are
  not granted on the basis of absolute population.
  Instead, electoral colleges are taken into
  account

El Gran Wyoming ha doblado su tamaño
6
Maps and cartograms of the 2004 US presidential
election results (4)

• and what about visualizing the results by
  county? (so USA Today did it!)

7
Maps and cartograms of the 2004 US presidential
election results (5)

• again it does not look the same if we use a
  cartogram

8
Maps and cartograms of the 2004 US presidential
election results (6)

• even less the same if we used non-linear blue
  and red combinations to introduce voting
  percentages (saturated at 70)

9
Maps and cartograms of the 2004 US presidential
election results (7)

• COMPARE!!!

10
Visualization
11
CRISP Methodology phases
12
CRISP Phases Data understanding
13
ReminderCRISP Tipology of DM problems
14
CRISP typology of DM problems examples
DESCRIPTION
15
CRISP typology of DM problems examples
clustering/segmentation (1)
16
CRISP typology of DM problems examples
clustering/segmentation (2)
17
Visualizing history
18
Once upon a time, circa 1600

• Michael van Langren, in 1644, displayed 12
  estimations of the longitude from Toledo to Rome
  This is, possibly, the earliest visualization of
  statistic data kept on record. A fuzzy arrow
  indicates the correct longitude (16o30') All
  estimations at the time were well off-mark. The
  word ROMA signals Langrens own estimation.

19
and reaching 1700

• Joseph Priestley generated this pioneering chart
  graphic display of v.i.p.s lives.
• (Source Joseph Priestly, A Chart of Biography,
  1765)

20
and speaking of the XVIII

• Immanuel Kant (1724-1804)
• Kant's very own philosophical "Copernican
  Revolution," it is the representation that makes
  the object possible rather than the object that
  makes the representation possible. This
  introduced the human mind as an active originator
  of experience rather than just a passive
  recipient of perception.

• After Kant, the old debate between rationalists
  and empiricists ended, and epistemology went in a
  new direction. After Kant, no discussion of
  reality or knowledge could take place without
  awareness of the role of the human mind in
  constructing reality and knowledge.

• Kant's approach is also of comparative interest
  because of the similar ancient Buddhist
  philosophical distinction between conditioned
  realities, which mostly means the world of
  experience, and unconditioned realities
  ("unconditioned dharmas"), which interestingly
  include space, which, for Kant, was a form
  imposed a priori on experience by the mind

21
introducing the industrial revolution

• William Playfair (XIX) created innovative
  graphics industrial / economic production time
  series and bar charts representing wheat prices,
  salaries, and monarchies along 250 years
  (Source Playfair, Letters on our agricultural
  distresses...)

22
Visualization
23
How do we process visualizations?Bases of visual
perception
24
PRINCIPLESvisual perception
Visualización
DATOS
MOTOR GRÁFICO
MODELO

• Pre-cognitive processing jigsaw pieces, or the
  interaction of visual elements.
• Visualization vs. mental model visual
  patterns, illusions and Gestalt laws.

25
Lets play!
26
PRINCIPLES visual perception example 1

• How many threes can you see?
• 4135761212521657124620246542024121642154154342157
  18413979421245401004687065487684617651168406540687
  436954387220254310520

• ... and in here?
• 4135761212521657124620246542024121642154154342157
  18413979421245401004687065487684617651168406540687
  436954387220254310520

27
PRINCIPLES visual perception example 2

• Abstract or figurative?

28
PRINCIPLES visual perception example 3
29
PRINCIPLES visual perception example 4

• Look at the words and say aloud what colour they
  are

AMARILLO AZUL NARANJA NEGRO ROJO VERDE VIOLETA A
MARILLO ROJO NARANJA VERDE NEGRO AZUL ROJO VIOLE
TA VERDE AZUL NARANJA

• Bipolarity conflict right hand side neo-cortex
  tries to name the colour but the left-nad side
  one insists on reading the word.

30
PRINCIPLESpre-cognitive processes

• Massive parallel processing of incoming visual
  stimuli, previous to their integration and
  subsequent conscious attention.
• Activity in cortical and subcortical areas, tuned
  to orientation, size, colour, movement and
  stereoscopic depth stimuli.

31
PRINCIPLESpre-cognitive elements
Movement
32
PRINCIPLESgrouping and perception

• Gestalt laws
• The properties of proximity and density are
  perceived as groupings.
• Similar objects belong to the same group
• Simetrical objects belong to the same group
• Objects within a perimeter belong to the same
  group
• Connected objects belong to the same group
• A collection of physical, biological,
  psychological or symbolic entities that creates a
  unified concept, configuration or pattern which
  is greater than the sum of its parts.
• Gestalt psychology is a theory of mind and brain
  that proposes that the operational principle of
  the brain is holistic, parallel, and analog, with
  self-organizing tendencies.

33
PRINCIPLESthe data mining visual cycle,
orVisual Exploratory Data Mining