The First History of the 2008 US Presidential Campaign Modeling - - PowerPoint PPT Presentation

▶
the first history of the 2008 us presidential campaign
SMART_READER_LITE
LIVE PREVIEW

The First History of the 2008 US Presidential Campaign Modeling - - PowerPoint PPT Presentation

Jan 07, 2024 217 likes •648 views

The First History of the 2008 US Presidential Campaign Modeling and Measuring Election Discourse Karl Grossner, PhD Candidate UC Santa Barbara Department of Geography 1 Geography and History Geography and history are different ways of

slide-1
SLIDE 1

The First History

  • f the

2008 US Presidential Campaign

Karl Grossner, PhD Candidate UC Santa Barbara Department of Geography

Modeling and Measuring Election Discourse

1

slide-2
SLIDE 2
slide-3
SLIDE 3

Geography and History

  • Geography and history are different ways of looking

at the world, but they are so closely related that neither one can afford to ignore or even neglect the

  • ther. (Baker, 2003 p.2)
  • Many more historical geographers than ‘geographical

historians’ but there is a ‘spatial turn’ under way.

  • HGIS is a growing field, as is ‘digital humanities’

Baker, A.R. (2003). Geography and History: Bridging the Divide. Cambridge: Cambridge University Press

slide-4
SLIDE 4

Towards a digital historical atlas

25 November 2008 spatial@ucsb brown bag

slide-5
SLIDE 5

Modeling and Measuring

Modeling Measuring

  • Variation in “issue aboutness”
  • f election discourse

– By region – By individual – By party – By media type

  • Representing complex social

events in databases, to support

– Discovery – Analysis – Reasoning – Mapping – Other visualizations

  • For historical scholarship

and education

– Digital historical atlases

25 November 2008 5 spatial@ucsb brown bag

slide-6
SLIDE 6

An Interactive Search Environment for Domain-Specific Knowledge Acquisition via Multimedia Data

IssueBrowser

Karl Grossner, Geography Jonathan Ventura, Computer Science Ben Adams, Computer Science Angus Forbes, Media Arts & Technology Monica Bulger, Education Swapna Joshi, Elec & Computer Engineering 6

slide-7
SLIDE 7

Graph Spatialization (2)

7

slide-8
SLIDE 8

Representing temporality in GIS

name year totalPop <5 (%) <18(%) female(%) Alameda 1900 130,197 4.8 21.1 46.3 Alameda 1910 246,131 5.9 27.3 50.9 name shape (polyline) t1 t2 Napoleon’s Russian campaign - advance

0101000020E6100000F2B5679604205

1812-06-23 1812-09-14 Napoleon’s Russian campaign - retreat

B56796042001010000F257C0E36BCF2

1812-10-19 1812-12-20 CA CA COUNTI ES COUNTI ES EVENTS EVENTS

(a radically oversimplified distinction)

name shape (polygon) t1 t2 Alameda

0101000020E6100000F2B5679604F05D

1867 1909 Alameda

00000F2B56796DC0E3620E6100001000

1910 2008 MODELING

25 November 2008 spatial@ucsb brown bag

slide-9
SLIDE 9

Napoleon’s Campaigns – In Fact

Armies Corps (#’s, nationalities) Generals Events Campaigns Battles (date, locs, outcomes) Fire River crossings Marches Flow (duration, #’s) States/conditions Bivouac Weather Temperature

9

MODELING

25 November 2008 spatial@ucsb brown bag

slide-10
SLIDE 10

Primitives of geographic knowledge

Golledge (1995) identity magnitude time location

  • ccurrence

In Nyerges, et al. (Eds) Proceedings of the NATO Advanced Research Workshop on Cognitive Aspects of Human-Computer Interaction for Geographic Information Systems, Palma de Mallorca, Spain, March 20-25. Boston:Kluwer Academic Publishers.

Golledge, R. G. (1995). Primitives of Spatial Knowledge. (measure) (temporality) name year totalPop <5 (%) <18(%) female(%) Alameda 1900 130,197 4.8 21.1 46.3 Alameda 1910 246,131 5.9 27.3 50.9 name shape (polyline) t1 t2 Napoleon’s Russian campaign - advance

0101000020E6100000F2B5679604205

1812-06-23 1812-09-14 Napoleon’s Russian campaign - retreat

B56796042001010000F257C0E36BCF2

1812-10-19 1812-12-20 needed: semantic reference system

  • cf. Kuhn, Raubal, Janowicz et al

identity

MODELING

25 November 2008 spatial@ucsb brown bag

slide-11
SLIDE 11

Westermann, U., and Jain, R. (2007). Toward a Common Event Model for Multimedia

  • Applications. IEEE Multimedia, 14(1), 19-29.

Eve nt

l o c a t i o n t i m e i d e n t i t y m a g n i t u d e c a u s a l i t y s t r u c t u r e MODELING

slide-12
SLIDE 12

To describe an event…

MODELING

25 November 2008 spatial@ucsb brown bag

slide-13
SLIDE 13

Ultimately modeling processes

  • Process (series of related events/activities)
  • beginning, formation, development, emergence
  • ending, collapse, destruction, disintegration
  • recovery
  • transition, change (of attributes, of identity), reshaping
  • expansion, growth (spatial or not)
  • contraction, diminishment
  • dispersal (abandon)
  • diffusion (spread from)
  • ascendency (of power, importance), rise, resurgence

MODELING

25 November 2008 spatial@ucsb brown bag

slide-14
SLIDE 14

Spatial Extents of Election 2008

(partial)

  • ther appearances

Republican Democrat debates

!

Democratic Debate

!

Republican Debate

speeches

#

Democrat

#

Republican

15 25 November 2008 spatial@ucsb brown bag

slide-15
SLIDE 15

Temporal Components of Election 2008

(partial)

16 25 November 2008 spatial@ucsb brown bag

slide-16
SLIDE 16

RDF/OWL

Classes Properties (relations)

 Actor

 Person  Group

 Information object

 linguistic object

 speech text  news report  blog entry

 Temporal entity

 Event

 political activity

 fundraiser

 Place

 administrative division (adm1)  populated place (ppl*)

  • isMemberOf/hasMember
  • isCreationOf/hasCreated
  • hadParticipant/participatedIn
  • hasPart/isPartOf
  • tookPlaceAt/witnessed

<subject,predicate,object> <Obama,isMemberOf,DemocraticParty> <DemConvention,hasPart,nomination tookPlaceAt,Denver.CO.US> tookPlaceAt domain: Event range: Place

history demands: attestedBy; assertedBy

MODELING

25 November 2008 spatial@ucsb brown bag

slide-17
SLIDE 17

CIDOC-CRM

E1 CRM Entity E2

  • Temporal Entity

E3

  • Condition State

E4

  • Period

E5

  • Event

E7

  • Activity

E63

  • Beginning of Existence

E67

  • Birth

E81

  • Transformation

E65

  • Creation

E66

  • Formation

E64

  • End of Existence

E77

  • Persistent Item

E70

  • Thing

E72

  • Legal Object

E71

  • Man-Made Thing

E24

  • -

Physical Man-Made Thing E22

  • -
  • Man-Made Object

E25

  • -
  • Man-Made Feature

E78

  • -
  • Collection

E28

  • -

Conceptual Object E73

  • -
  • Information Object

E29

  • -
  • -

Design or Procedure E31

  • -
  • -

Document E33

  • -
  • -

Linguistic Object E36

  • -
  • -

Visual Item E55

  • -
  • Type

E39

  • Actor

E74

  • Group

E21

  • Person

E41

  • Appellation

E42

  • Identifier

E44

  • Place Appellation

E49

  • Time Appellation

E75

  • Conceptual Object Appellation

E82

  • Actor Appellation

Property Name

Entity – Domain Entity - Range P1 is identified by (identifies) E1 CRM Entity E41 Appellation P2 has type (is type of) E1 CRM Entity E55 Type P3 has note E1 CRM Entity E62 String P4 has time-span (is time-span of) E2 Temporal Entity E52 Time-Span P5 consists of (forms part of) E3 Condition State E3 Condition State P7 took place at (witnessed) E4 Period E53 Place P8 took place on or within (witnessed) E4 Period E19 Physical Object P9 consists of (forms part of) E4 Period E4 Period P10 falls within (contains) E4 Period E4 Period P12

  • ccurred in the presence of

E5 Event E77 Persistent Item P11

  • had participant

E5 Event E39 Actor P16

  • used specific object

E7 Activity E70 Thing P25

  • moved

E9 Move E19 Physical Object P31

  • has modified

E11 Modification E24 Physical Man-Made Thing P33

  • used specific technique

E7 Activity E29 Design or Procedure P92

  • brought into existence

E63 Beginning of Existence E77 Persistent Item P93

  • took out of existence

E64 End of Existence E77 Persistent Item P15 was influenced by (influenced) E7 Activity E1 CRM Entity P19 was intended use of (was made for) E7 Activity E71 Man-Made Thing P20 had specific purpose E7 Activity E7 Activity P32 used general technique E7 Activity E55 Type P43 has dimension (is dimension of) E70 Thing E54 Dimension P44 has condition (condition of) E18 Physical Thing E3 P45 consists of (is incorporated in) E18 Physical Thing E57 Material P46 is composed of (forms part of) E18 Physical Thing E18 Physical Thing

85 class declarations 148 property declarations MODELING

slide-18
SLIDE 18

CIDOC-CRM

CIDOC Conceptual Reference Model (ISO 21127:2006); Doerr, Crofts, Gill, Stead, Stiff

an ontology of cultural heritage information

MODELING

25 November 2008 spatial@ucsb brown bag

slide-19
SLIDE 19

CIDOC-CRM

  • Rudimentary reasoning:

– sub-classes and sub-properties (isA inheritance) – property holds for domain with valid values in range – inverse and symmetric properties – cardinality constraints

P1 is identified by (identifies) E1 CRM Entity E41 Appellation P2 has type (is type of) E1 CRM Entity E55 Type P11 had participant E5 Event E39 Actor P15 was influenced by (influenced) E7 Activity E1 CRM Entity P20 had specific purpose E7 Activity E7 Activity Property Name Entity – Domain Entity - Range P127 has broader term (has narrower term) E55 Type E55 Type

MODELING

25 November 2008 spatial@ucsb brown bag

slide-20
SLIDE 20

US Election 2008

  • Ontology developed in RDF/OWL using Protégé
  • On a CIDOC-CRM scaffold

– some problems, e.g. activity vs. action; overly complex time constructs

  • Instantiated in a RDBMS (PostgreSQL)

– a model isn’t useful if not used: compatible with GIS – dissimilar logics: relational algebra vs. description logic – numerous RDBMS tools: procedural languages; 3rd-party functions (spatial, tsearch, recursion); custom data types (e.g. temporal)

MODELING

25 November 2008 spatial@ucsb brown bag

slide-21
SLIDE 21

A more general model

MODELING

slide-22
SLIDE 22

Measuring Election Discourse

  • Candidate speeches

– 227 Democrat (126 Obama) – 121 Republican (77 McCain)

  • Debate transcripts

– 21 Democratic, 19 Republican, 4 post-primaries

  • Wire service news reports (geography-free? AP, Reuters, etc.)

– ~12,000 between Nov 2006 and now

  • Regional newspaper reports

– ~22,600 reports from 36 newspapers in 9 census regions from Jan 2008

  • Blog text since the conventions

– Liberal: Huffington Post, Daily Kos – Conservative: Hot Air, Right Wing News

  • Candidate issue statements

– 577 statements in 58 issue categories  34 issues MEASURING

25 November 2008 spatial@ucsb brown bag

slide-23
SLIDE 23

Issue Signatures (1)

Service to America Thank you. It's good to be back in Meridian. As you might know, I was once a flight instructor here at the air field named for my grandfather during my long past and misspent youth. And it's always good to be in Mississippi, which you could call my ancestral

  • home. Generations of McCains were born and raised in Carroll County, on land that had been in our family since 1848. The last

McCain to live on the property, which the family called Teoc, was my grandfather's brother, Joe McCain. I spent a couple summers here as a young boy, and enjoyed it immensely. I had never had a permanent address because my father's naval career required us to move frequently. But here, in the care of my very likeable Uncle Joe, I could imagine, with a little envy, what it must have been like for the McCains who came before me to be so connected to one place; to be part of a community and a landscape as well as a family. 24

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-24
SLIDE 24

Issue Signatures (2)

sum of the dot product

  • product of the euclidean ‘distances’

http://www.miislita.com/information-retrieval-tutorial/cosine-similarity-tutorial.html 25

MEASURING

slide-25
SLIDE 25

Issue Signatures (3)

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-26
SLIDE 26

Results (1)

0.1 0.2 0.3 0.4 0.5 0.6 abortion economy education energy environment healthcare immigration reform conflict values

389 Speeches -- All Candidates summarized issue bins

E N Cen E S Cen Mid Atl Mtn N Eng Pacific S Atl W N Cen W S Cen random

economy

  • “the economy”
  • spending
  • taxes
  • poverty
  • social security
  • the middle class

environment

  • environment
  • global warming

conflict

  • global image
  • iraq war
  • war on terror
  • foreign policy
  • defense policy

values

  • faith and politics
  • families
  • service

reform

  • political reform
  • voting reform

energy

  • energy policy
  • energy independence

0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 i5_0 i5_1 i5_2 i5_3 i5_5 i5_6 i7_0 i7_1 i8_0 i8_1 i16_0 i18_0 i19_0 i20_0 i21_0 Pacific Mtn W N Cen W S Cen E N Cen E S Cen N Eng Mid Atl S Atl random

economy energy environment conflict

cosine similarity cosine similarity

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-27
SLIDE 27

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-28
SLIDE 28

Results (3)

cosine similarity cosine similarity

0.05 0.1 0.15 0.2 0.25 0.3

81 Democratic Swing State Speeches

NV CO MO OH PA NC VA FL random 0.05 0.1 0.15 0.2 0.25 0.3

55 Republican Swing State Speeches

NV CO MO OH PA NC VA FL random

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-29
SLIDE 29

Swing state issue emphasis by party/region

MEASURING

slide-30
SLIDE 30

The Final Four

McCain/Palin & Obama/Biden

0.0000 0.0200 0.0400 0.0600 0.0800 0.1000 0.1200 0.1400 jm sp bo jb random

cosine similarity

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-31
SLIDE 31

Appalachia

pct McCain vote

0.128 - 0.419 0.419 - 0.540 0.540 - 0.627 0.627 - 0.713 0.713 - 0.844

29 Elmira Star-Gazette 82 Binghamton Press & Sun-Bulletin 535 Erie Times-News 125 Scranton Times-Tribune 503 Charleston Gazette 299 Knoxville News Sentinel 80 Asheville Citizen-Times 137 Winston-Salem Journal 21 Huntsville Times

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-32
SLIDE 32

MEASURING Salient terms (stemmed) from newspaper articles, extracted and weighted with Latent Dirichlet Allocation (LDA) clustering algorithm

29 Elmira Star-Gazette 82 Binghamton Press & Sun-Bulletin 535 Erie Times-News 125 Scranton Times-Tribune 503 Charleston Gazette 299 Knoxville News Sentinel 80 Asheville Citizen-Times 137 Winston-Salem Journal 21 Huntsville Times

Northern Appalachia Southern Appalachia

slide-33
SLIDE 33

classical MDS 34  2 Democratic Convention Republican Convention MEASURING

slide-34
SLIDE 34

34 dimensions  2 w/MDS

77 speeches 44 Obama, B. 3 Biden 2 Obama, M. 20 McCain 8 Palin MEASURING

slide-35
SLIDE 35

Travels through (semantic) space

Nov 4 Sep 9 – Oct 2 Oct 26 – Nov 3 Oct 15 – Oct 25 Oct 3 – Oct 13 Sep 19 – Oct 10 Oct 16 – Oct 20 Nov 4

77 speeches 44 Obama, B. 3 Biden 2 Obama, M. 20 McCain 8 Palin MEASURING

slide-36
SLIDE 36

Flesch-Kincaid (1)

10 20 30 40 50 60 70 80 6 8 10 12 14 16 18

Ease of Understandability Grade Level

2008 Candidate Speeches Flesch-Kincaid Readability

  • bama

biden mccain palin

Rudolf Flesch (1948); A new readability yardstick, Journal of Applied Psychology, Vol. 32, pp. 221-233 Kincaid, J. P.; Fishburne, R. P., Jr.; Rogers, R. L.; and Chissom, B. S. (1975); Derivation of new readability formulas (Automated Readability Index, Fog Count and Flesch Reading Ease Formula) for Navy enlisted personnel, Research Branch Report 8-75, Millington, TN: Naval Technical Training, U. S. Naval Air Station, Memphis, TN

MEASURING

slide-37
SLIDE 37

Flesch-Kincaid (2)

10 20 30 40 50 60 70 80 90 3 5 7 9 11 13

Reading Ease Grade Level

Convention Speeches Flesch-Kincaid Readability

Democratic (27 speeches) Republican (18 speeches) Various

BY GRADE LEVEL

Gettysburg Address Obama’s “Race Speech” Roosevelt 3rd Inaugural Lincoln 2nd Inaugural Kennedy Inaugural GW Bush 2nd Inaugural Clinton 2nd Inaugural M L King “I Have a Dream” Nixon Resignation

Al Gore Tom Ridge Cindy McCain McCain (1)

  • B. Clinton

MEASURING

25 November 2008 spatial@ucsb brown bag

McCain (2) Obama

slide-38
SLIDE 38

Flesch-Kincaid (3)

Average grade level Democrat 10.3 Republican 10.6

2 4 6 8 10 12 Pacific Mtn W N Cen W S Cen E N Cen E S Cen N Eng Mid Atl S Atl

Campaign Speech Grade Level by Region

Democratic Republican

MEASURING

25 November 2008 spatial@ucsb brown bag

slide-39
SLIDE 39

Conclusions (Election 2008)

  • There is geographic variation in issue aboutness insofar as

candidates are concerned

  • Democratic and Republican speech occupies distinct,

marginally overlapping regions of semantic space

  • Candidates discuss issues more than news reports do
  • The Obama campaign was in fact rhetorically strategic and

disciplined (and/or sitting on a lead)

  • Some of my conceits were wrong (cf. McCain speechwriters’

reading level)

  • Some were right
  • The internet has complicated geography

25 November 2008 spatial@ucsb brown bag

slide-40
SLIDE 40

Further Questions

  • What are the actual issues discussed, e.g.

change, leadership, Wright & Ayers, VP competency

– analyze news reports, blogs

  • Does an ontology-based data model have

value for focused geo-historical analyses?

– (not just interoperability, sharing, encyclopedic systems)

25 November 2008 spatial@ucsb brown bag

slide-41
SLIDE 41

Comments, questions? And thanks to NSF IGERT Grant # DGE-0221713

25 November 2008 spatial@ucsb brown bag