Fear and loathing on the social campaign trail

Stuart Shulman
Stuart ShulmanCEO at Texifter
Fear and Loathing on the
Social Campaign Trail
Dr. Stuart Shulman
@stuartwshulman
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Part One
Political Fear and Loathing
in United States History
“The distinguishing thing about
the paranoid style is not that its
exponents see conspiracies or
plots here and there in history,
but that they regard a vast' or
gigantic' conspiracy as the
"motive force" in historical
events...The paranoid spokesman
sees the fate of this conspiracy in
apocalyptic terms--he traffics in
the birth and death of whole
worlds, whole political orders,
whole systems of human values.
He is always manning the
barricades of civilization.”
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Emergent properties found in a very well read texts,
such as the character type “extremist agent of the law”
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Part Two
The Social Campaign Trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Part Three
Previous Work on Methods & Tools
The NRC emotion lexicon is a list of
words and their associations with eight
emotions (anger, fear, anticipation,
trust, surprise, sadness, joy, and
disgust) and two sentiments (negative
and positive).
The First 150 of 1476 NRC Fear Words
Coding Fear Words
Coding Fear Words
Fear and loathing on the social campaign trail
- MIT Professor Eric von Hippel
“This is really the biggest paradigm shift in innovation since the Industrial Revolution”
Crowdsourcing brings widely distributed
wisdom to process of text analysis
Coder Count Unit Count
13 9
12 47
11 27
10 80
9 155
8 67
7 27
6 8
5 17
4 8
3 474
2 957
1 1757
Crowds Create New Filtering Options
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Measurement of Coder Agreement
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Defined Search - Incompetence
Fear and loathing on the social campaign trail
Clearly Fear?
Adjudicating Fear Coding
Clearly Fear?
Ideology
Custom Classifier Histogram
Filtering Using a Classifier Histogram
Part Four
New Work on Methods & Tools
First Person Political Fear Tweets
Version 1 Gnip PowerTrack Rule
i (hate OR fear OR loathe OR despise OR
dislike OR abhor OR aversion OR afraid OR
scared OR dismay OR dread OR horror OR
alarm OR frightened OR frightful OR
horrified OR terrified) "american politics"
lang:en -is:retweet
First Person Tweet Collection
Version 2 Gnip PowerTrack Rule
(hate OR fear OR loathe OR despise OR
dislike OR abhor OR aversion OR afraid OR
scared OR dismay OR dread OR horror OR
alarm OR frightened OR frightful OR
horrified OR terrified) "american politics"
lang:en -is:retweet
First Person Tweet Collection
First Person Tweet Collection
Version 3 Gnip PowerTrack Rule
("i fear" OR "i am afraid" OR "i'm scared"
OR "i am scared" OR "i am worried" OR
"i'm worried" OR "i dread" OR "i am
horrified" OR "i worry" OR "i feel afraid"
OR "i feel scared" OR "i am terrified" OR "i
feel terrified" OR "i feel worried" OR
"worries me" OR "scares me" OR
"frightens me" OR "horrifies me" OR
"terrifies me") (trump) lang:en -is:retweet
Version 4 Gnip PowerTrack Rule
("i fear" OR "i am afraid" OR "i'm scared" OR
"i am scared" OR "i am worried" OR "i'm
worried" OR "i dread" OR "i am horrified"
OR "i worry" OR "i feel afraid" OR "i feel
scared" OR "i am terrified" OR "i feel
terrified" OR "i feel worried" OR "worries
me" OR "scares me" OR "frightens me" OR
"horrifies me" OR "terrifies me") (libtard OR
democrat OR liberal) lang:en -is:retweet
First Person Tweet Collection
Version 5 Gnip PowerTrack Rule
("i fear" OR "i am afraid" OR "i'm scared" OR
"i am scared" OR "i am worried" OR "i'm
worried" OR "i dread" OR "i am horrified" OR
"i worry" OR "i feel afraid" OR "i feel scared"
OR "i am terrified" OR "i feel terrified" OR "i
feel worried" OR "worries me" OR "scares
me" OR "frightens me" OR "horrifies me" OR
"terrifies me") (libtards OR democrats OR
liberals) lang:en -is:retweet
First Person Tweet Collection
The Archives in the Pilot
Deduplication without Retweets
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Checking Inter-Rater Reliability
•We conducted four reliability checks
•Datasets were 200, 200, 100, and 200 items
•We used between 6 & 12 coders
•Fleiss’ kappa = .76, .91, .80, and .85
Checking Validity
•We conducted regular validity checks
•Thousands of observations were validated
•Very few invalid observations overall
•Invalid observations not used for training
•Better quality training data
•The “gold standard”
•Better understanding in the 50+ page codebook
• We rank coders all the time.
• CoderRank is the notion that for any
annotation task, simple to complex,
there is a range of human aptitude.
• A small number of coders are fantastic.
• Surprisingly small at times.
• A larger number is awful.
• Especially for hard tasks!
• Most are average.
• ~65-85% valid.
CoderRankSM
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
Fear and loathing on the social campaign trail
5732814
Fear and loathing on the social campaign trail
Dr. Stuart W. Shulman
Founder & CEO, Texifter, LLC
Editor Emeritus, Journal of Information Technology & Politics
stu@texifter.com
@stuartwshulman
Thank-you for having me!
1 of 88

Recommended

Citations & Paraphrasing by
Citations & ParaphrasingCitations & Paraphrasing
Citations & ParaphrasingCristy Bolton
437 views14 slides
Hvordan være relevant i en alder dominert av digitale medier by
Hvordan være relevant i en alder dominert av digitale medierHvordan være relevant i en alder dominert av digitale medier
Hvordan være relevant i en alder dominert av digitale medierNebojsha Mihajlovski
306 views41 slides
AI-based rumor & fake news detection algorithm on Twitter by
AI-based rumor & fake news detection algorithm on TwitterAI-based rumor & fake news detection algorithm on Twitter
AI-based rumor & fake news detection algorithm on TwitterMeeyoung Cha
790 views16 slides
What Can We Learn from the Unabomber?: Nothing. by
What Can We Learn from the Unabomber?: Nothing.What Can We Learn from the Unabomber?: Nothing.
What Can We Learn from the Unabomber?: Nothing.Peter Ludlow
3.7K views101 slides
Extreme Abuse 101: Expanding Our Therapeutic Container to Better Serve Our Cl... by
Extreme Abuse 101: Expanding Our Therapeutic Container to Better Serve Our Cl...Extreme Abuse 101: Expanding Our Therapeutic Container to Better Serve Our Cl...
Extreme Abuse 101: Expanding Our Therapeutic Container to Better Serve Our Cl...Staci Sprout, LICSW, CSAT
400 views72 slides
Agression in humans and non humans by
Agression in humans and non humansAgression in humans and non humans
Agression in humans and non humansAlice Palmer
455 views16 slides

More Related Content

Similar to Fear and loathing on the social campaign trail

Animal Experimentation by
Animal ExperimentationAnimal Experimentation
Animal ExperimentationAllison Thompson
2 views21 slides
LIAR by
LIARLIAR
LIARGordon Fowkes
208 views5 slides
The Rights Of Animal Rights Essay by
The Rights Of Animal Rights EssayThe Rights Of Animal Rights Essay
The Rights Of Animal Rights EssayOnline Paper Writer Singapore
3 views20 slides
The Architecture of Understanding by
The Architecture of UnderstandingThe Architecture of Understanding
The Architecture of UnderstandingPeter Morville
2.2K views58 slides
Galileo Research: The State of the (Dis)Union by
Galileo Research: The State of the (Dis)UnionGalileo Research: The State of the (Dis)Union
Galileo Research: The State of the (Dis)UnionMeg Stagaard
141 views18 slides

Similar to Fear and loathing on the social campaign trail(6)

More from Stuart Shulman

Fear and Loathing on the Social Campaign Trail by
Fear and Loathing on the Social Campaign TrailFear and Loathing on the Social Campaign Trail
Fear and Loathing on the Social Campaign TrailStuart Shulman
128 views88 slides
Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi... by
Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi...Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi...
Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi...Stuart Shulman
170 views80 slides
Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase! by
Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase!Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase!
Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase!Stuart Shulman
217 views13 slides
CoderRank: Creating Gold Standards by
CoderRank: Creating Gold StandardsCoderRank: Creating Gold Standards
CoderRank: Creating Gold StandardsStuart Shulman
245 views22 slides
Text Analytics for Social Data Using DiscoverText & Sifter by
 Text Analytics for Social Data Using DiscoverText & Sifter Text Analytics for Social Data Using DiscoverText & Sifter
Text Analytics for Social Data Using DiscoverText & SifterStuart Shulman
809 views60 slides
Text Analytics for Social Data Using DiscoverText & Sifter by
Text Analytics for Social Data Using DiscoverText & SifterText Analytics for Social Data Using DiscoverText & Sifter
Text Analytics for Social Data Using DiscoverText & SifterStuart Shulman
336 views60 slides

More from Stuart Shulman(18)

Fear and Loathing on the Social Campaign Trail by Stuart Shulman
Fear and Loathing on the Social Campaign TrailFear and Loathing on the Social Campaign Trail
Fear and Loathing on the Social Campaign Trail
Stuart Shulman128 views
Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi... by Stuart Shulman
Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi...Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi...
Text Analytics: From Colored Pens and Crumbly Papers to Custom Machine Classi...
Stuart Shulman170 views
Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase! by Stuart Shulman
Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase!Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase!
Texifter Presentation at Boston New Technology’s #BNT77 Startup Showcase!
Stuart Shulman217 views
CoderRank: Creating Gold Standards by Stuart Shulman
CoderRank: Creating Gold StandardsCoderRank: Creating Gold Standards
CoderRank: Creating Gold Standards
Stuart Shulman245 views
Text Analytics for Social Data Using DiscoverText & Sifter by Stuart Shulman
 Text Analytics for Social Data Using DiscoverText & Sifter Text Analytics for Social Data Using DiscoverText & Sifter
Text Analytics for Social Data Using DiscoverText & Sifter
Stuart Shulman809 views
Text Analytics for Social Data Using DiscoverText & Sifter by Stuart Shulman
Text Analytics for Social Data Using DiscoverText & SifterText Analytics for Social Data Using DiscoverText & Sifter
Text Analytics for Social Data Using DiscoverText & Sifter
Stuart Shulman336 views
Sifting Social Data: Word Sense Disambiguation Using Machine Learning by Stuart Shulman
Sifting Social Data: Word Sense Disambiguation Using Machine LearningSifting Social Data: Word Sense Disambiguation Using Machine Learning
Sifting Social Data: Word Sense Disambiguation Using Machine Learning
Stuart Shulman1.5K views
CAQDAS 2014 Pecha Kucha - Stuart Shulman by Stuart Shulman
CAQDAS 2014 Pecha Kucha - Stuart ShulmanCAQDAS 2014 Pecha Kucha - Stuart Shulman
CAQDAS 2014 Pecha Kucha - Stuart Shulman
Stuart Shulman485 views
Measuring reliability and validity in human coding and machine classification by Stuart Shulman
Measuring reliability and validity in human coding and machine classificationMeasuring reliability and validity in human coding and machine classification
Measuring reliability and validity in human coding and machine classification
Stuart Shulman844 views
Technology for Citizen Voices by Stuart Shulman
Technology for Citizen VoicesTechnology for Citizen Voices
Technology for Citizen Voices
Stuart Shulman591 views
DiscoverText: Tools for Text by Stuart Shulman
DiscoverText: Tools for TextDiscoverText: Tools for Text
DiscoverText: Tools for Text
Stuart Shulman1.6K views
Citizen Voices in a Networked Age of #BigData by Stuart Shulman
Citizen Voices in a Networked Age of #BigDataCitizen Voices in a Networked Age of #BigData
Citizen Voices in a Networked Age of #BigData
Stuart Shulman916 views
DiscoverText Product Overview by Stuart Shulman
DiscoverText Product OverviewDiscoverText Product Overview
DiscoverText Product Overview
Stuart Shulman399 views
Importing bulk outlook email into DiscoverText - the .pst file upload by Stuart Shulman
Importing bulk outlook email into DiscoverText - the .pst file uploadImporting bulk outlook email into DiscoverText - the .pst file upload
Importing bulk outlook email into DiscoverText - the .pst file upload
Stuart Shulman377 views
Future of text analysis forrester briefing by Stuart Shulman
Future of text analysis   forrester briefingFuture of text analysis   forrester briefing
Future of text analysis forrester briefing
Stuart Shulman459 views

Recently uploaded

SOCO 9.pdf by
SOCO 9.pdfSOCO 9.pdf
SOCO 9.pdfSocioCosmos
6 views1 slide
"Mastering Social Media Marketing: A Guide to Fremont's Local Influence and C... by
"Mastering Social Media Marketing: A Guide to Fremont's Local Influence and C..."Mastering Social Media Marketing: A Guide to Fremont's Local Influence and C...
"Mastering Social Media Marketing: A Guide to Fremont's Local Influence and C...Embtel Solutions
24 views19 slides
Soco 7.pdf by
Soco 7.pdfSoco 7.pdf
Soco 7.pdfSocioCosmos
9 views1 slide
PDF.pdf by
PDF.pdfPDF.pdf
PDF.pdfoliverumr
11 views1 slide
Soco 11 (2).pdf by
Soco 11 (2).pdfSoco 11 (2).pdf
Soco 11 (2).pdfSocioCosmos
6 views1 slide
digital marketing by
digital marketing digital marketing
digital marketing mdZafar18
5 views1 slide

Recently uploaded(10)

Fear and loathing on the social campaign trail

  • 1. Fear and Loathing on the Social Campaign Trail Dr. Stuart Shulman @stuartwshulman
  • 10. Part One Political Fear and Loathing in United States History
  • 11. “The distinguishing thing about the paranoid style is not that its exponents see conspiracies or plots here and there in history, but that they regard a vast' or gigantic' conspiracy as the "motive force" in historical events...The paranoid spokesman sees the fate of this conspiracy in apocalyptic terms--he traffics in the birth and death of whole worlds, whole political orders, whole systems of human values. He is always manning the barricades of civilization.”
  • 19. Emergent properties found in a very well read texts, such as the character type “extremist agent of the law”
  • 29. Part Two The Social Campaign Trail
  • 34. Part Three Previous Work on Methods & Tools
  • 35. The NRC emotion lexicon is a list of words and their associations with eight emotions (anger, fear, anticipation, trust, surprise, sadness, joy, and disgust) and two sentiments (negative and positive).
  • 36. The First 150 of 1476 NRC Fear Words
  • 40. - MIT Professor Eric von Hippel “This is really the biggest paradigm shift in innovation since the Industrial Revolution” Crowdsourcing brings widely distributed wisdom to process of text analysis
  • 41. Coder Count Unit Count 13 9 12 47 11 27 10 80 9 155 8 67 7 27 6 8 5 17 4 8 3 474 2 957 1 1757 Crowds Create New Filtering Options
  • 44. Measurement of Coder Agreement
  • 51. Defined Search - Incompetence
  • 58. Filtering Using a Classifier Histogram
  • 59. Part Four New Work on Methods & Tools
  • 60. First Person Political Fear Tweets
  • 61. Version 1 Gnip PowerTrack Rule i (hate OR fear OR loathe OR despise OR dislike OR abhor OR aversion OR afraid OR scared OR dismay OR dread OR horror OR alarm OR frightened OR frightful OR horrified OR terrified) "american politics" lang:en -is:retweet First Person Tweet Collection
  • 62. Version 2 Gnip PowerTrack Rule (hate OR fear OR loathe OR despise OR dislike OR abhor OR aversion OR afraid OR scared OR dismay OR dread OR horror OR alarm OR frightened OR frightful OR horrified OR terrified) "american politics" lang:en -is:retweet First Person Tweet Collection
  • 63. First Person Tweet Collection Version 3 Gnip PowerTrack Rule ("i fear" OR "i am afraid" OR "i'm scared" OR "i am scared" OR "i am worried" OR "i'm worried" OR "i dread" OR "i am horrified" OR "i worry" OR "i feel afraid" OR "i feel scared" OR "i am terrified" OR "i feel terrified" OR "i feel worried" OR "worries me" OR "scares me" OR "frightens me" OR "horrifies me" OR "terrifies me") (trump) lang:en -is:retweet
  • 64. Version 4 Gnip PowerTrack Rule ("i fear" OR "i am afraid" OR "i'm scared" OR "i am scared" OR "i am worried" OR "i'm worried" OR "i dread" OR "i am horrified" OR "i worry" OR "i feel afraid" OR "i feel scared" OR "i am terrified" OR "i feel terrified" OR "i feel worried" OR "worries me" OR "scares me" OR "frightens me" OR "horrifies me" OR "terrifies me") (libtard OR democrat OR liberal) lang:en -is:retweet First Person Tweet Collection
  • 65. Version 5 Gnip PowerTrack Rule ("i fear" OR "i am afraid" OR "i'm scared" OR "i am scared" OR "i am worried" OR "i'm worried" OR "i dread" OR "i am horrified" OR "i worry" OR "i feel afraid" OR "i feel scared" OR "i am terrified" OR "i feel terrified" OR "i feel worried" OR "worries me" OR "scares me" OR "frightens me" OR "horrifies me" OR "terrifies me") (libtards OR democrats OR liberals) lang:en -is:retweet First Person Tweet Collection
  • 66. The Archives in the Pilot
  • 76. Checking Inter-Rater Reliability •We conducted four reliability checks •Datasets were 200, 200, 100, and 200 items •We used between 6 & 12 coders •Fleiss’ kappa = .76, .91, .80, and .85
  • 77. Checking Validity •We conducted regular validity checks •Thousands of observations were validated •Very few invalid observations overall •Invalid observations not used for training •Better quality training data •The “gold standard” •Better understanding in the 50+ page codebook
  • 78. • We rank coders all the time. • CoderRank is the notion that for any annotation task, simple to complex, there is a range of human aptitude. • A small number of coders are fantastic. • Surprisingly small at times. • A larger number is awful. • Especially for hard tasks! • Most are average. • ~65-85% valid. CoderRankSM
  • 88. Dr. Stuart W. Shulman Founder & CEO, Texifter, LLC Editor Emeritus, Journal of Information Technology & Politics stu@texifter.com @stuartwshulman Thank-you for having me!