Alexandra Olteanu, Sarah Vieweg and Carlos Castillo: "What to Expect When the Unexpected Happens: Social Media Communications Across Crises" In CSCW 2015, 14-18 March in Vancouver, Canada. ACM Press.
4. What are the similarities and differences in Twitter
communications that take place during different crisis
events, according to speciďŹc characteristics of such
events?
5. What are the similarities and differences in Twitter
communications that take place during different crisis
events, according to speciďŹc characteristics of such
events?
6. What are the similarities and differences in Twitter
communications that take place during different crisis
events, according to speciďŹc characteristics of such
events?
18. Step 2: Content Dimensions
â Informativeness
Informative:
useful information, situational information, etc.
Not informative:
prayers, trolling, spam, humor, rumor, off-topic, etc.
â Source of information
Primary sources
Eyewitness: citizen reporters, local individuals, direct experience
Secondary & tertiary sources
Government: authorities, police & ďŹre services, public institutions
NGOs: non-proďŹt org., non-governmental org., faith-based org.
Business: comercial org., enterprises, for-proďŹt corporation
Media: news org., journalists, news media!
Outsiders: remote crowd, non-locals, sympathizers
â Type of information
!
!
!
!
!
!
!
!
!
19. Step 2: Content Dimensions
â Informativeness
Informative:
useful information, situational information, etc.
Not informative:
prayers, trolling, spam, humor, rumor, off-topic, etc.
â Source of information
Primary sources
Eyewitness: citizen reporters, local individuals, direct experience
Secondary & tertiary sources
Government: authorities, police & ďŹre services, public institutions
NGOs: non-proďŹt org., non-governmental org., faith-based org.
Business: comercial org., enterprises, for-proďŹt corporation
Media: news org., journalists, news media!
â Type of information
!
!
!
!
!
!
20. Step 2: Content Dimensions
â Informativeness
Informative:
useful information, situational information, etc.
Not informative:
prayers, trolling, spam, humor, rumor, off-topic, etc.
â Source of information
Primary sources
Eyewitness: citizen reporters, local individuals, direct experience
Secondary & tertiary sources
Government: authorities, police & ďŹre services, public institutions
NGOs: non-proďŹt org., non-governmental org., faith-based org.
Business: comercial org., enterprises, for-proďŹt corporation
Media: news org., journalists, news media!
Outsiders: remote crowd, non-locals, sympathizers
â Type of information
!
!
!
!
!
!
!
!
!
21. Step 2: Content Dimensions
â Informativeness
Informative:
useful information, situational information, etc.
Not informative:
prayers, trolling, spam, humor, rumor, off-topic, etc.
â Source of information
Primary sources
Eyewitness: citizen reporters, local individuals, direct experience
Secondary & tertiary sources
Government: authorities, police & ďŹre services, public institutions
NGOs: non-proďŹt org., non-governmental org., faith-based org.
Business: comercial org., enterprises, for-proďŹt corporation
Media: news org., journalists, news media!
Outsiders: remote crowd, non-locals, sympathizers
â Type of information
22. Step 2: Content Dimensions
â Informativeness
Informative:
useful information, situational information, etc.
Not informative:
prayers, trolling, spam, humor, rumor, off-topic, etc.
â Source of information
Primary sources
Eyewitness: citizen reporters, local individuals, direct experience
Secondary & tertiary sources
Government: authorities, police & ďŹre services, public institutions
NGOs: non-proďŹt org., non-governmental org., faith-based org.
Business: comercial org., enterprises, for-proďŹt corporation
Media: news org., journalists, news media!
Outsiders: remote crowd, non-locals, sympathizers
â Type of information
!
!
!
!
!
!
!
!
!
23. Step 2: Content Dimensions
â Informativeness
Informative:
useful information, situational information, etc.
Not informative:
prayers, trolling, spam, humor, rumor, off-topic, etc.
â Source of information
Primary sources
Eyewitness: citizen reporters, local individuals, direct experience
Secondary & tertiary sources
Government: authorities, police & ďŹre services, public institutions
NGOs: non-proďŹt org., non-governmental org., faith-based org.
Business: comercial org., enterprises, for-proďŹt corporation
Media: news org., journalists, news media!
Outsiders: remote crowd, non-locals, sympathizers
â Type of information
!
!
!
!
!
!
!
!
!
Affected individuals!
casualties; people missing, found, trapped, seen; reports about self
Infrastructure & utilities!
road closures, collapsed structure, water sanitation, services
Donations & volunteering
requesting help, proposing relief, relief coordination, shelter needed
Caution & advice!
predicting or forecasting, instructions to handle certain situations
Sympathy & emotional support!
thanks, gratitude, prayers, condolences, emotion-related
Other useful information!
meta-discussions, ďŹood level, wind, visibility, weather
24. Step 3: Data Collection
â Twitter base-sample*
â 2012 & 2013
â ~1% random sample of Twitter public stream
â ~130+ million tweets per month
â Keyword-based searches
â proper names of affected location
â manila ďŹoods, boston bombings, #newyork derailment
â proper names of meteorological phenomena
â sandy hurricane, typhoon yolanda
â promoted hashtags
â #SafeNow, #RescuePH, #ReliefPH
â 26 crisis events
â 14 countries and 8 languages
â 12 different hazard types
â earthquake, wildďŹre, ďŹoods, bombings, shootings, etc.
!
â 15 instantaneous crises
â 15 diffused crises
* https://archive.org/details/twitterstream
25. Step 3: Data Collection
â Twitter base-sample*
â 2012 & 2013
â ~1% random sample of Twitter public stream
â ~130+ million tweets per month
â Keyword-based searches
â proper names of affected location
â manila ďŹoods, boston bombings, #newyork derailment
â proper names of meteorological phenomena
â sandy hurricane, typhoon yolanda
â promoted hashtags
â #SafeNow, #RescuePH, #ReliefPH
â 26 crisis events
â 14 countries and 8 languages
â 12 different hazard types
â earthquake, wildďŹre, ďŹoods, bombings, shootings, etc.
!
â 15 instantaneous crises
â 15 diffused crises
* https://archive.org/details/twitterstream
26. Step 3: Data Collection
â Twitter base-sample*
â 2012 & 2013
â ~1% random sample of Twitter public stream
â ~130+ million tweets per month
â Keyword-based searches
â proper names of affected location
â manila ďŹoods, boston bombings, #newyork derailment
â proper names of meteorological phenomena
â sandy hurricane, typhoon yolanda
â promoted hashtags
â #SafeNow, #RescuePH, #ReliefPH
â 26 crisis events
â 14 countries and 8 languages
â 12 different hazard types
â earthquake, wildďŹre, ďŹoods, bombings, shootings, etc.
!
â 15 instantaneous crises
â 15 diffused crises
* https://archive.org/details/twitterstream
27. Step 3: Data Collection
â Twitter base-sample*
â 2012 & 2013
â ~1% random sample of Twitter public stream
â ~130+ million tweets per month
â Keyword-based searches
â proper names of affected location
â manila ďŹoods, boston bombings, #newyork derailment
â proper names of meteorological phenomena
â sandy hurricane, typhoon yolanda
â promoted hashtags
â #SafeNow, #RescuePH, #ReliefPH
â 26 crisis events
â 14 countries and 8 languages
â 12 different hazard types
â earthquake, wildďŹre, ďŹoods, bombings, shootings, etc.
!
â 15 instantaneous crises
â 15 diffused crises
* https://archive.org/details/twitterstream
28. Step 3: Data Collection
â Twitter base-sample*
â 2012 & 2013
â ~1% random sample of Twitter public stream
â ~130+ million tweets per month
â Keyword-based searches
â proper names of affected location
â manila ďŹoods, boston bombings, #newyork derailment
â proper names of meteorological phenomena
â sandy hurricane, typhoon yolanda
â promoted hashtags
â #SafeNow, #RescuePH, #ReliefPH
â 26 crisis events
â 14 countries and 8 languages
â 12 different hazard types
â earthquake, wildďŹre, ďŹoods, bombings, shootings, etc.
!
â 15 instantaneous crises
â 15 diffused crises
* https://archive.org/details/twitterstream
29. Step 3: Data Collection
â Twitter base-sample*
â 2012 & 2013
â ~1% random sample of Twitter public stream
â ~130+ million tweets per month
â Keyword-based searches
â proper names of affected location
â manila ďŹoods, boston bombings, #newyork derailment
â proper names of meteorological phenomena
â sandy hurricane, typhoon yolanda
â promoted hashtags
â #SafeNow, #RescuePH, #ReliefPH
â 26 crisis events
â 14 countries and 8 languages
â 12 different hazard types
â earthquake, wildďŹre, ďŹoods, bombings, shootings, etc.
!
â 15 instantaneous crises
â 15 diffused crises
* https://archive.org/details/twitterstream
30. Step 4: Data Annotation
â ~1000 tweets per crisis
â Informativeness
â Source of information
â Type of information
!
â Crowdsource workers from the affected countries
31. Step 5: Data Analysis
â Content types/sources vs. crisis dimensions
â Interplay between types and sources
â Crisis similarity
â Temporal aspects
32. What: Types of Information
Infrastructure and utilities: 7% on average (min. 0%, max. 22%)!
most prevalent in diffused crises, in particular during ďŹoods!
Caution and advice: 10% on average (min. 0%, max. 34%)!
least prevalent in instantaneous & human-induced disasters!
Donations and volunteering: 10% on average (min. 0%, max. 44%)!
most prevalent in natural disasters
Costa Rica earthquakeâ12
Manila ďŹoodsâ13
Singapore hazeâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Australia bushďŹreâ13
Italy earthquakesâ12
Sardinia ďŹoodsâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Typhoon Yolandaâ13
Colorado ďŹoodsâ13
Guatemala earthquakeâ12
Colorado wildďŹresâ12
Bohol earthquakeâ13
NY train crashâ13
Boston bombingsâ13
LA airport shootingsâ13
West Texas explosionâ13
Russia meteorâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Venezuela reďŹneryâ12
Glasgow helicopter crashâ13
Spain train crashâ13
Brazil nightclub ďŹreâ13
0
10
20
30
40
50
60
70
80
90
100
Caution & Advice
Affected Ind.
Infrast. & Utilities
Donat. & Volun.
Sympathy
Other Useful Info.
33. Affected individuals: 20% on average (min. 5%, max. 57%)!
most prevalent in human-induced, focalized & instantaneous crises!
Sympathy and emotional support: 20% on average (min. 3%, max. 52%)!
most prevalent in instantaneous crises!
Other useful information: 32% on average (min. 7%, max. 59%)!
least prevalent in diffused crises
Costa Rica earthquakeâ12
Manila ďŹoodsâ13
Singapore hazeâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Australia bushďŹreâ13
Italy earthquakesâ12
Sardinia ďŹoodsâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Typhoon Yolandaâ13
Colorado ďŹoodsâ13
Guatemala earthquakeâ12
Colorado wildďŹresâ12
Bohol earthquakeâ13
NY train crashâ13
Boston bombingsâ13
LA airport shootingsâ13
West Texas explosionâ13
Russia meteorâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Venezuela reďŹneryâ12
Glasgow helicopter crashâ13
Spain train crashâ13
Brazil nightclub ďŹreâ13
0
10
20
30
40
50
60
70
80
90
100
Caution & Advice
Affected Ind.
Infrast. & Utilities
Donat. & Volun.
Sympathy
Other Useful Info.
What: Types of Information
34. Who: Sources of Information
Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado ďŹoodsâ13
Colorado wildďŹresâ12
Bohol earthquakeâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Venezuela reďŹneryâ12
West Texas explosionâ13
Sardinia ďŹoodsâ13
Spain train crashâ13
Guatemala earthquakeâ12
Brazil nightclub ďŹreâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Russia meteorâ13
Lac Megantic train crashâ13
Typhoon Yolandaâ13
Savar building collapseâ13
NY train crashâ13
0
10
20
30
40
50
60
70
80
90
100
Eyewitness
Government
NGOs
Business
Media
Outsiders
Business: 2% on average (min. 0%, max. 9%)!
most prevalent in diffused crises!
NGOs: 4% on average (min. 0%, max. 17%)!
most prevalent in natural disasters, in particular during typhoons & ďŹoods!
Government: 5% on average (min. 1%, max. 13%)!
most prevalent in natural, progressive & diffused crises
35. Eyewitness accounts: 9% on average (min. 0%, max. 54%)!
most prevalent in progressive & diffused crises!
Outsiders: 38% on average (min. 3%, max. 65%) !
least in the Singapore Haze crisis!
Traditional & Internet media: 42% on average (min. 18%, max. 77%) !
most prevalent in instantaneous crises, which make the âbreaking newsâ
Who: Sources of Information
Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado ďŹoodsâ13
Colorado wildďŹresâ12
Bohol earthquakeâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Venezuela reďŹneryâ12
West Texas explosionâ13
Sardinia ďŹoodsâ13
Spain train crashâ13
Guatemala earthquakeâ12
Brazil nightclub ďŹreâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Russia meteorâ13
Lac Megantic train crashâ13
Typhoon Yolandaâ13
Savar building collapseâ13
NY train crashâ13
0
10
20
30
40
50
60
70
80
90
100
Eyewitness
Government
NGOs
Business
Media
Outsiders
36. Events Similarity: Information Types
Savar building collapseâ13
LA airport shootingsâ13
NY train crashâ13
Russia meteorâ13
Colorado wildďŹresâ12
Guatemala earthquakeâ12
Glasgow helicopter crashâ13
West Texas explosionâ13
Lac Megantic train crashâ13
Venezuela reďŹneryâ12
Bohol earthquakeâ13
Boston bombingsâ13
Brazil nightclub ďŹreâ13
Spain train crashâ13
Manila ďŹoodsâ13
Alberta ďŹoodsâ13
Philipinnes ďŹoodsâ12
Typhoon Yolandaâ13
Costa Rica earthquakeâ12
Singapore hazeâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Queensland ďŹoodsâ13
Colorado ďŹoodsâ13
Sardinia ďŹoodsâ13
Typhoon Pabloâ12
lower similarity
37. Events Similarity: Information Types
Savar building collapseâ13
LA airport shootingsâ13
NY train crashâ13
Russia meteorâ13
Colorado wildďŹresâ12
Guatemala earthquakeâ12
Glasgow helicopter crashâ13
West Texas explosionâ13
Lac Megantic train crashâ13
Venezuela reďŹneryâ12
Bohol earthquakeâ13
Boston bombingsâ13
Brazil nightclub ďŹreâ13
Spain train crashâ13
Manila ďŹoodsâ13
Alberta ďŹoodsâ13
Philipinnes ďŹoodsâ12
Typhoon Yolandaâ13
Costa Rica earthquakeâ12
Singapore hazeâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Queensland ďŹoodsâ13
Colorado ďŹoodsâ13
Sardinia ďŹoodsâ13
Typhoon Pabloâ12
lower similarity
38. Events Similarity: Information Types
Savar building collapseâ13
LA airport shootingsâ13
NY train crashâ13
Russia meteorâ13
Colorado wildďŹresâ12
Guatemala earthquakeâ12
Glasgow helicopter crashâ13
West Texas explosionâ13
Lac Megantic train crashâ13
Venezuela reďŹneryâ12
Bohol earthquakeâ13
Boston bombingsâ13
Brazil nightclub ďŹreâ13
Spain train crashâ13
Manila ďŹoodsâ13
Alberta ďŹoodsâ13
Philipinnes ďŹoodsâ12
Typhoon Yolandaâ13
Costa Rica earthquakeâ12
Singapore hazeâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Queensland ďŹoodsâ13
Colorado ďŹoodsâ13
Sardinia ďŹoodsâ13
Typhoon Pabloâ12
Dominated by
natural, diffused
and progressive
lower similarity
39. Events Similarity: Information Types
Savar building collapseâ13
LA airport shootingsâ13
NY train crashâ13
Russia meteorâ13
Colorado wildďŹresâ12
Guatemala earthquakeâ12
Glasgow helicopter crashâ13
West Texas explosionâ13
Lac Megantic train crashâ13
Venezuela reďŹneryâ12
Bohol earthquakeâ13
Boston bombingsâ13
Brazil nightclub ďŹreâ13
Spain train crashâ13
Manila ďŹoodsâ13
Alberta ďŹoodsâ13
Philipinnes ďŹoodsâ12
Typhoon Yolandaâ13
Costa Rica earthquakeâ12
Singapore hazeâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Queensland ďŹoodsâ13
Colorado ďŹoodsâ13
Sardinia ďŹoodsâ13
Typhoon Pabloâ12
Dominated by
natural, diffused
and progressive
Dominated by
human-induced,
focalized and
instantaneous
lower similarity
40. Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado wildďŹresâ12
Colorado ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Typhoon Yolandaâ13
Brazil nightclub ďŹreâ13
Russia meteorâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Bohol earthquakeâ13
Venezuela reďŹneryâ12
Sardinia ďŹoodsâ13
West Texas explosionâ13
NY train crashâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Guatemala earthquakeâ12
Spain train crashâ13
lower similarity
Events Similarity: Information Sources
41. Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado wildďŹresâ12
Colorado ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Typhoon Yolandaâ13
Brazil nightclub ďŹreâ13
Russia meteorâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Bohol earthquakeâ13
Venezuela reďŹneryâ12
Sardinia ďŹoodsâ13
West Texas explosionâ13
NY train crashâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Guatemala earthquakeâ12
Spain train crashâ13
lower similarity
Events Similarity: Information Sources
42. Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado wildďŹresâ12
Colorado ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Typhoon Yolandaâ13
Brazil nightclub ďŹreâ13
Russia meteorâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Bohol earthquakeâ13
Venezuela reďŹneryâ12
Sardinia ďŹoodsâ13
West Texas explosionâ13
NY train crashâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Guatemala earthquakeâ12
Spain train crashâ13
lower similarity
Dominated by instantaneous,
focalized and human-induced
Events Similarity: Information Sources
43. Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado wildďŹresâ12
Colorado ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Typhoon Yolandaâ13
Brazil nightclub ďŹreâ13
Russia meteorâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Bohol earthquakeâ13
Venezuela reďŹneryâ12
Sardinia ďŹoodsâ13
West Texas explosionâ13
NY train crashâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Guatemala earthquakeâ12
Spain train crashâ13
lower similarity
Dominated by instantaneous,
focalized and human-induced
Events Similarity: Information Sources
44. Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado wildďŹresâ12
Colorado ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Typhoon Yolandaâ13
Brazil nightclub ďŹreâ13
Russia meteorâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Bohol earthquakeâ13
Venezuela reďŹneryâ12
Sardinia ďŹoodsâ13
West Texas explosionâ13
NY train crashâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Guatemala earthquakeâ12
Spain train crashâ13
lower similarity
Dominated by instantaneous,
focalized and human-induced
Events Similarity: Information Sources
45. Singapore hazeâ13
Philipinnes ďŹoodsâ12
Alberta ďŹoodsâ13
Manila ďŹoodsâ13
Italy earthquakesâ12
Australia bushďŹreâ13
Colorado wildďŹresâ12
Colorado ďŹoodsâ13
Queensland ďŹoodsâ13
Typhoon Pabloâ12
Typhoon Yolandaâ13
Brazil nightclub ďŹreâ13
Russia meteorâ13
Boston bombingsâ13
Glasgow helicopter crashâ13
Bohol earthquakeâ13
Venezuela reďŹneryâ12
Sardinia ďŹoodsâ13
West Texas explosionâ13
NY train crashâ13
Costa Rica earthquakeâ12
LA airport shootingsâ13
Savar building collapseâ13
Lac Megantic train crashâ13
Guatemala earthquakeâ12
Spain train crashâ13
lower similarity
Dominated by instantaneous,
focalized and human-induced
Dominated by
natural, diffused
and progressive
Events Similarity: Information Sources
65. 12h 24h 36h 48h ⌠several days
peak
Instantaneous
Temporal Distribution: Sources
66. 12h 24h 36h 48h ⌠several days
peak
Eyewitness
Instantaneous
Temporal Distribution: Sources
67. 12h 24h 36h 48h ⌠several days
peak
Eyewitness
Outsiders
Instantaneous
Temporal Distribution: Sources
68. 12h 24h 36h 48h ⌠several days
peak
Eyewitness
Outsiders
Media
Instantaneous
Temporal Distribution: Sources
69. 12h 24h 36h 48h ⌠several days
peak
Eyewitness
Government
Outsiders
Media
Instantaneous
Temporal Distribution: Sources
70. 12h 24h 36h 48h ⌠several days
peak
Eyewitness
Government
NGOs
Outsiders
Media
Instantaneous
Temporal Distribution: Sources
71. 12h 24h 36h 48h ⌠several days
peak
Eyewitness
Government
Business
NGOs
Outsiders
Media
Instantaneous
Temporal Distribution: Sources
72. Take-Away
Twitter is a medium through which the nuance of events is
ampliďŹed; yet, when looking at the same data at a higher-
level we see commonalities and patterns.
!
!
â Download all our collections from crisislex.org
Thanks!
Karl Aberer
@ChatoX
@velofemme
Patrick Meier
Questions?
@o_saja
74. Temporal Distribution
ProgressiveInstantaneous
⢠Start: when the event occurs
⢠High volumes of tweets right
after onset
⢠Start: when the hazard is detected
⢠High volumes around the peak of
the event (e.g., affects a densely
populated area, high economic
damage)
75. Goal: Retrieve comprehensive collections of event-
related messages
⢠Keyword-based sampling: â¨
#sandy, #bostonbombings, #qldďŹood
⢠Location-based sampling: â¨
tweets geo-tagged in disaster area
Problem: Create event collections without too many off-
topic tweets
Recall(% of on-crisis tweets retrieved)
Precision
(on-crisis%ofretrievedtweets)
DesiredKW-based
Geo-based
Lexicon-based
ICWSM 2014, Olteanu et al.
EfďŹcient Data Collection
76. CrisisLex
API limits
⢠rigid query language
⢠limited volumes
Laconic language
Challenges
damage
affected people
people displaced
donate blood
text redcross
stay safe
crisis deepens
evacuated
toll raises
âŚâŚ
ICWSM 2014, Olteanu et al.
80. Association-Rules
â Diffused events have more than average caution &
advice messages
â valid for 24/26 events
!
â Human-induced & accidental events have less than
average eyewitness accounts
â valid for 21/26 events
81. Informativeness &
Content Redundancy
â Informativeness
â Crisis-related: 89% on average (min. 64%, max. 100%)
â Informative (from crisis-related): 69% on average (min.
44%, max. 92%)
!
â Redundancy
â NGOs & Government
â top 3 messages account for 20%-22% of messages
!
â Caution & Advice and Infrastructure & Utilities
â top 3 messages account for 12%-14% of messages