This document discusses social network analysis (SNA) and big data research methods based on social media and big data. It provides an overview of SNA, including definitions of key graph theory concepts like nodes, edges, centrality measures, components, cliques and triads. It also discusses how SNA can be used to analyze information diffusion on social networks, identify influential users, and study how conflicts may propagate through social ties between users. Real-world examples are provided on how SNA has been applied to topics like politics, celebrity networks, and terrorist cells.

1. SOCIAL NETWORK
ANALYSIS
METODE PENELITIAN BERBASIS
SOCIAL MEDIA DAN BIG DATA
Ismail Fahmi, Ph.D.
Director Media Kernels Indonesia (Drone Emprit)
Lecturer at the University of Islam Indonesia
Ismail.fahmi@gmail.com
KULIAH TAMU FISHUM UIN JOGJA
5 NOVEMBER 2020

2. 2
1992 – 1997 S1, Teknik Elektro, ITB
2003 – 2004 S2, Information Science, Universitas Groningen, Belanda
2004 – 2009 S3, Information Science, Universitas Groningen, Belanda
2000 – 2003 Inisiator IndonesiaDLN (Digital Library Network pertama di Indonesia)
Mengembangkan Ganesha Digital Library (GDL)
Mendirikan Knowledge Management Research Group (KMRG) ITB
Membangun Digital Library ITB
2009 – Sekarang Engineer di Weborama, Perusahaan berbasis big data (Paris/Amsterdam)
2014 – Sekarang Founder PT. Media Kernels Indonesia, a Drone Emprit Company
2015 – Sekarang Konsultan Perpustakaan Nasional, Inisiator Indonesia OneSearch
2017 – Sekarang Dosen Tetap Magister Teknik Informatika Universitas Islam Indonesia
Ismail Fahmi, Ph.D.
Ismail.fahmi@gmail.com
Lahir: Bojonegoro, 1974
Founder Media Kernels Indonesia

3. AGENDA
• Tentang Drone Emprit
• Understanding Social Media Data Sources
• Analytics Workflow and Settings
• Twitter Data Crawling
• Social Network Analysis
• Cliques and Components
• Triads, Network Density, and Conflict
• Viralitas dan Difusi Informasi
• SNA dan Twitter
• Contoh
• Konflik dan Viralitas
• Viralitas Omnibus Law
• Bot dan SNA
3

4. TENTANG DRONE EMPRIT

5. ABOUT PT. MEDIA KERNELS INDONESIA
PT. Media Kernels Indonesia is focused on harnessing Natural
Language Processing (NLP) technologies to provide innovative
solutions in big data, text mining and insight discovery for
knowledge-based institutions.
Our products:
• Media Kernels (aka Drone Emprit), a media monitoring and
analytics tool.
• Fact Miner, an information extraction and visualization tool for
unstructured text.
5
FactMiner

6. TENTANG DRONE EMPRIT ACADEMIC
Drone Emprit Academic adalah sebuah
sistem big data yang menangkap dan
menganalisis percakapan di media sosial
khususnya Twitter, yang dikembangkan
oleh PT Media Kernels Indonesia, dan
bekerjasama dengan Universitas Islam
Indonesia untuk penyediaan layanannya.
Drone Emprit menggunakan layanan API
(Applications Programming Interface) dari
Twitter untuk menangkap percakapan
secara semi realtime melalui metode
streaming.
6

7. DRONE EMPRIT ACADEMIC
FREE SOCIAL MEDIA (TWITTER) DATA ANALYTICS
7

8. JOIN DRONE EMPRIT ACADEMIC
HTTPS://DEA.UII.AC.ID
8

9. HOW IT WORKS
9
STEPS:
• Registration
• Propose keywords
• Analysis and publication
Dashboard
Access
REQUIREMENTS:
• Publish their analysis for public
using any medium at least 1
publication every 2 months.
USERS
• Students
• Researchers
• Lecturers
• Journalists
• Blogger
• Hoax buster
Admin

10. TOPICS BASED ON SDGS
(SUSTAINABLE DEVELOPMENT GOALS)
10

11. DRONE EMPRIT ACADEMIC DASHBOARD
11

12. UNDERSTANDING SOCIAL MEDIA
DATA SOURCES

13. BIG DATA – BIG GROWTH
13

14. PENGGUNA TWITTER DI INDONESIA
NAIK DARI 27% (2018) MENJADI 56% (2020)
14
27%
52%
56%
2018
2019
2020

15. 15

16. 16
MILLENIALGEN Z

17. WORKFLOW & SETTING

18. WORKFLOW
18
Data
Population
(backtrack)
Analysis Visualization
Sentiment analysis,
opinion analysis, bot
analysis, demography
analysis, etc
Keyword dan filter Social Network Analysis,
Tree Map, Geolocation
Map, Trends, etc.

19. RESEARCH QUESTIONS &
KEYWORD SETTINGS
• Saat memulai analisis, yang perlu dibuat
adalah: Research Questions.
• Insight apa saja yang ingin didapat dari analisis.
• Bisa menggunakan kerangka 5W + 1H dalam
menyusun pertanyaan.
• Kemudian atur setting kata kunci dan filter
untuk mendapatkan “populasi percakapan”
seakurat dan selengkap mungkin.
• Percakapan di media sosial yang terkumpul
menggunakan setting ini tidak boleh
mengandung noise (percakapan tidak relevan)
terlalu banyak.
• Tidak boleh juga terlalu sedikit, banyak yang
hilang, karena ingin akurat.
19

20. DATA IS VERY EXPENSIVE TO GET
20

21. TANTANGAN: DATA CRAWLING

22. KEYWORD & FILTER
22
Sensus, sp2020, ..
Sensus
penduduk
Keyword à Twitter
Sumber daya yang sangat terbatas
Filter à Data Lake
Membatasi hasil pencarian

23. BERHEMAT DENGAN KEYWORD
(DIAGRAM VENN)
23
Sensus, sp2020, ..
Sensus penduduk

24. Data Lake
Keywords
KEYWORD + FILTER
24
Twits
twit
twit
twit
Filters

25. FULL ARCHIVE (BERBAYAR)
25

26. DATA IS EXPENSIVE
100 USD PER 50.000 TWEETS
26

27. CONTOH: FREE TWITTER SEARCH
27
History: 7 days
Start search
100% results

28. METODE PENELITIAN BIG DATA DAN SNA
• Merupakan metode
gabungan:
• Kuantitatif
• Kualitatif
28

29. SOCIAL NETWORK ANALYSIS

30. REFERENSI BUKU

31. SOCIAL NETWORK ANALYSIS
FOR STARTUPS
31

32. SOCIAL NETWORK ANALYSIS
HANDBOOK
32

33. GRAPH THEORY
• Dyad: unit terkecil dari SNA (Social Network Analysis)
• Terdiri dari:
• Node
• Link
• Node
33
- node
- vertex
- edge
- link
- relationship

34. 1-MODE GRAPH
• Menghubungkan tipe node yang sama, misal:
• Orang dengan orang
• Organisasi dengan organisasi
• Kata dengan kata
• dst
34
orang orang

35. 2-MODE GRAPH
• Menghubungkan 2 tipe node yang berbeda, misal:
• Orang dengan organisasi
• Orang dengan point of interest
• Orang dengan hashtags
• dst
35
orang Point of interest
Ancol

36. MULTI-MODE GRAPH
• Menghubungkan:
• Orang dengan:
• Orang
• Organisasi
• Point of interest
• Hashtags
• etc
36
orang Point of interest
Ancol
orang

37. SINGLE VERB
• Pada umumnya link dalam sebuah network menggunakan verb
yang sama, misal: like.
• Di media sosial, link ini misalnya:
• Twitter: retweet, mention, follows.
• Facebook: friends, like, reply to.
37

38. MULTI VERBS
• Dalam dunia nyata, verb dalam link bisa beragam, misal:
• Like
• Study
• Fight
38

39. LINK BISA MEMILIKI VALUE
• Skala Likert bisa digunakan, misal:
• 0. Don't know
1. Strongly dislike
2. Dislike
3. Neither dislike nor like 4. Like
5. Strongly like
• Atau dalam sentiment analysis:
• Positive (hijau)
• Negative (merah)
• Neutral (abu-abu)
39

40. ADJACENCY MATRICES
• Social network bisa direpresentasikan secara matematis
menggunakan matrik:
• Cell dengan angka: 0,1 (ada atau tidaknya link)
• Cell dengan angka: 0,1,2,3,4,3 (nilai dari link)
• Kelemahan: akan banyak cell dengan nilai 0
40

41. EDGE LIST
• Solusi dari banyaknya 0 dalam adjacency matrices, menampilkan
link sebagai sebuah urutan (list).
• Hanya menampilkan link yang punya value saja.
41

42. GRAPH TRAVERSALS & DISTANCES
• Dalam sebuah graph, kita bisa
berjalan dari satu node ke
node lainnya.
• Jarak perjalanan bisa dekat,
paling dekat, atau jauh;
tergantung dari banyaknya
node yang harus dilewati.
• Misal, dari node 0 ke 7, paling
dekat melalui node 5.
42

43. DIJKSTRA’S ALGORITHM
• For a given vertex it finds
the lowest cost path to all
other vertices, where
“cost” is determined by
summing edge weights.
• In graphs where edge
weights correspond to
distance (in unweighted
graphs the weights are
assumed to be one) the
found path is the shortest.
• Contoh, dari node 1 ke 4:
• Shortest path: 1,4
• Lowest path (Dijkstra):
1,0,3,6,4
43

44. CENTRALITY
• Centrality merupakan sebuah metode untuk mengukur power dan
influence (dari individual/orang/node).
• Cara mengukur centrality:
• Degree centrality
• Betweenness centrality
• Closeness centrality
• Eigenvector centrality
• PageRank
44
https://cambridge-intelligence.com/social-network-analysis/

45. DEGREE CENTRALITY
• The degree centrality measure
finds nodes with the highest
number of links to other nodes
in the network (out link). Nodes
with a high degree centrality have
the best connections to those
around them – they might be
influential, or just strategically
well-placed.
45

46. BETWEENNESS CENTRALITY
• Nodes with a high betweenness
centrality score are the ones that
most frequently act as ‘bridges’
between other nodes. They form
the shortest pathways of
communication within the network.
• Usually this would indicate
important gatekeepers of
information between groups.
46

47. CONTOH BETWEENNESS
47
Betweenness (Bridge)

48. CLOSENESS CENTRALITY
• Definition: Closeness centrality scores
each node based on their ‘closeness’ to
all other nodes in the network.
• What it tells us: This measure calculates
the shortest paths between all nodes,
then assigns each node a score based on
its sum of shortest paths.
• When to use it: For finding the
individuals who are best placed to
influence the entire network most quickly.
• A bit more detail: Closeness centrality
can help find good ‘broadcasters’, but in a
highly-connected network, you will often
find all nodes have a similar score. What
may be more useful is using Closeness to
find influencers in a single cluster.
48

49. EIGENVECTOR CENTRALITY
• Definition: Like degree centrality,
EigenCentrality measures a node’s
influence based on the number of links
it has to other nodes in the network.
EigenCentrality then goes a step
further by also taking into account how
well connected a node is, and how
many links their connections have, and
so on through the network.
• What it tells us: By calculating the
extended connections of a node,
EigenCentrality can identify nodes with
influence over the whole network, not
just those directly connected to it.
• When to use it: EigenCentrality is a
good ‘all-round’ SNA score, handy for
understanding human social networks,
but also for understanding networks
like malware propagation.
49

50. PAGERANK
• PageRank identifies
important nodes by
assigning each a score
based upon its number of
incoming links (its
‘indegree’). These links
are weighted depending
on the relative score of its
originating node.
50

51. PEMANFAATAN CENTRALITY
• Menemukan:
• Celebritis
• Tukang gossip
• Communication bottlenecks
• Communication bridges
51

52. CLIQUES AND COMPONENTS

53. COMPONENTS
• Component subgraphs
(or simply components)
are portions of the
network that are
disconnected from
each other.
• A subgraph is a subset
of the nodes of a
network, and all of the
edges linking these
nodes.
53
Nested Components

54. ISLANDS IN THE NET
• The giant component gets
split up into smaller
components, and areas with
the strongest amount of
retweeting activity (subcores)
become their own
components that can be
analyzed separately.
54

55. EGO NETWORK
• Ego Netowork is a subnetwork that is centered on a certain
node.
55

56. TRIAD
• A triad is simply three nodes interlinked in some way.
56

57. TRIADS AND TERRORISTS
• Al Qaeda cells were often
sequestered in safe-houses
during training and
preparation for terrorist
attacks.
• The most common factor
driving them was the social
ties within their cell. Most
started as friends, colleagues,
or relatives—and were drawn
closer by bonds of friendship,
loyalty, solidarity and trust,
and rewarded by a powerful
sense of belonging and
collective identity.
57

58. TRIADS AND KPOPPERS
• Kpoppers drawn closer by bonds of friendship, loyalty, solidarity and trust,
and rewarded by a powerful sense of belonging and collective identity.
58

59. TRIADS IN POLITICS
• The network in the figure is
built by looking at the joint
political statements and
agreements made by the
countries and republics of
the Caucasus, Russia,
Turkey, EU, and the United
States, and clearly shows
the different governing
styles of Russia and the
West.
• The Russia-centric side of
the network shows a system
rich in structural holes.
Russia is truly in charge, and
lateral ties between
peripheral actors are almost
non-existent.
59

60. CLIQUES
• A clique is defined as a
maximal complete
subgraph of a given
graph—i.e., a group of
people where everybody is
connected directly to
everyone else.
• A clique consists of several
overlapping closed triads,
and inherits many of the
culture-generating, and
amplification properties of
closed triads.
60

61. CONTOH CLIQUE DAN COMPONENT
• DE sering menganalisis
SNA dan membedahnya ke
dalam “clusters”, yang
sebenarnya disebut
“component.”
• Contoh SNA di samping,
setidaknya terbentuk atas
beberapa component.
Tampak sebuah clique dari
3 akun yang terindikasi
robot.
61
Component
Component
Component
Clique

62. TRIADS, NETWORK DENSITY, AND
CONFLICT

63. FRIENDSHIP AND CONFLICT: RULES
1. Friend of my friend is
my friend (close a
structural hole)
2. Enemy of my friend is
my enemy (achieve a
balanced triad)
• Friend of my enemy is my
enemy
• Enemy of my enemy is my
friend
63

64. CLOSING TRIADS AND VIRALITY
• If an open triad A→B→C is detected,
with some probability we will also
add a link A→ C.
• We are still adding links randomly, so
at first, the network will grow linearly.
• At a certain point, a critical mass of
connection has been created, and
every new connection is likely to
create an open triad.
• This open triad is then closed by Rule
1, which may in turn create more
open triads, which then get closed,
and so on.
• In a sense, the network passes from a
linear growth to an exponential
growth. It goes viral!
64

65. CONFLICT PROPAGATION
• Conflict is introduced in the network at a
constant probability, by changing a single
friendship tie into an enemy tie.
• In this simple example, a network consisting
of 4 closed triads is struck by a conflict on a
single edge.
• Triad A−B−C becomes unbalanced due to a
conflict between B and C; thus A is forced to
take sides in the conflict by choosing to
remain friends with either B or C, at random.
• Adding conflict to the A−C edge forces
another triad (A−C−D) to become
unbalanced, thus drawing agent D into the
conflict.
• If agent D then chooses to isolate C from the
rest of the network, the propagation of the
conflict can be stopped.
• However, if instead it separates from A, this
will cause the conflict to propagate further
and destroy more links.
65

66. CONFLICT AND NETWORK DENSITY
• Having more ties increases an
agent’s probability of forming even
more ties, but also increases the
probability that a conflict between
two agents will spread throughout
the network.
• As a result, network density no
longer grows to near 100%; but
instead, once it reaches a second
critical mass value, conflicts
become more prominent and bring
the network density back down.
66

67. VIRALITAS DAN DIFUSI INFORMASI

68. DIFUSI INFORMASI
68

69. ANATOMI VIRALITAS SEBUAH PESAN
• It largely depends on who sees it and what
they do with it afterwards.
• At first, the only people that see it are in your
immediate ego network—i.e., your friends.
• As such, growth in the number of views is very
slow; it is linear over time, meaning that views
happen at a constant rate. In this case, we can
describe the number of views as a Poisson
process, where every act of viewing your cat
video is independent of every other act.
• At the end, the total number of views you
would get would be mathematically related to
the number of followers you have—your
degree centrality.
• But if something else happens and your
friends retweet the video to their friends (and
so on, exposing more and more people to it),
your cat is suddenly in the most popular video
on YouTube.
69

70. CRITICAL MASS
• If the transition from linear propagation
to exponential (viral) growth indeed
depends on triadic closure (i.e., “friend
of a friend is my friend”),† then the
critical mass of connections can be
estimated by measuring the probability
that a randomly created link from A to B
will form one or more open triads with
other nodes.
• The transition from linear growth (where
connections are added one by one) to
going viral happens somewhere around
7% density —i.e., if 7% of people in the
intended audience adopt a meme,
retweet a video, join a site, etc., the rest
will follow shortly in a viral wave.
• This was the “magic number” that
pushed Facebook over the edge at
Harvard.
70

71. WIKINOMICS OF CRITICAL MASS
• Let us imagine that the cost of participation
remains constant.
• To an early adopter, the cost is very real and
the benefit is yet to materialize. As more
people join the network, the benefit rises as
a function of number of connections, making
the new idea or meme or social network site
an easier sell every step of the way.
• Critical mass will be reached when the
benefits of using the product start
outweighing the cost; at this point, each
connection will in turn breed more
connections, further increasing the benefit—
while the cost still remains constant.
• If the cost/benefit crossover is never
reached, it may not matter how many
connections are in the network—eventually
the network will fail.
71

72. CONTENT IS (STILL) KING: EFEK PESAN
• Relevance
Do I care at all? (And its variant, Saliency—do I care right now?)
• Resonance
Does the content of the message mesh with what I already
believe in?
• Severity
How good or bad is the content of the message?
• Immediacy
Does this message demand an immediate action? Combined
with Severity, what is the consequence of inaction?
• Certainty
Does the effect of this message cause certain pain or pleasure
or is the chance low?
• Source
Who did this message come from and do I trust this person?
Has it been corrobo- rated by someone else?
• Entertainment value
Is it funny? A good read?
72

73. SOCIAL NETWORK ANALYSIS DAN TWITTER

74. RELASI RETWEET
74

75. RELASI: RETWEET (AGREEMENT) VS MENTION
(DISCUSSION)
75

76. RELASI RETWEET
76

77. RELASI MENTION
77

78. SNA TAGAR (WHO SAYS WHAT)
#DiperkosaNegara
GEN Z / K-POPPERS
#TurunkanJokowi
OPOSISI
#ReformasiDikorupsi
AKTIVIS
#PercayaLangkahJokowi
#SayaBersamaJokowi
#JokowiMendengarRakyat
#KitaDukungJokowi
PRO PEMERINTAH
TRIK GIVE AWAY

79. CONTOH KONFLIK DAN VIRALITAS

80. VOLUME
#GejayanMemanggil2
#MahasiswaPelajarAnarkis

81. TREN
#GejayanMemanggil2
#MahasiswaPelajarAnarkis

82. DISINFORMASI BISA DICIPTAKAN DENGAN CARA MURAH MERIAH:
MELALUI AKUN-AKUN ANONIM, GIVEAWAY, DAN BUZZER
1) SEBUAH DISINFORMASI BISA DIMULAI DARI
SEBUAH AKUN ANONIM ATAU GIVEAWAY,
YANG DIAMPLIFIKASI OLEH AKUN-AKUN
SEPERTI BOT, AGAR #TAGAR CEPAT
TRENDING
2) KARENA TRENDING, #TAGAR MUDAH
DILIHAT OLEH AKUN-AKUN NATURAL, YANG
TURUT MENGAMPLIFIKASI JIKA COCOK,
ATAU MENCOUNTER JIKA TIDAK SETUJU
3) DI PENGHUJUNG HARI, AKUN GIVEAWAY YANG
MENGINISASI #TAGAR TIDAK LAGI DOMINAN, KARENA
PERCAKAPAN SUDAH DIAMBIL ALIH OLEH AKUN-AKUN
NATURAL DAN VIRAL
#MahasiswaPelajarAnarkis
AKUN ANONIM
AKUN ANONIM
AKUN ANONIM
AKUN MAHASISWA
AKUN MAHASISWA
AKUN MAHASISWA
TAGAR #MAHASISWAPELAJARANARKIS DICIPTAKAN UNTUK
MENGACAUKAN NARASI DALAM GERAKAN MAHASISWA
#GEJAYANMEMANGGIL2, MELALUI AKUN GIVEAWAY, ANONIM
DAN BUZZER

83. DAMPAK SEBARAN DISINFORMASI:
NARASI PENTING TERTUTUPI OLEH DISINFORMASI
KONTRA MAHASISWA PRO MAHASISWA
#GejayanMemanggil2
#MahasiswaPelajarAnarkis
DISINFORMASI SERING SENGAJA DICIPTAKAN
UNTUK MEMBENTUK OPINI PUBLIK DEMI
KEPENTINGAN TERTENTU, DENGAN ATAU TANPA
BANTUAN AGEN KOMUNIKASI PUBLIK
STARTER BUZZERS
TARGET
TARGET
TARGET
TARGET
KETIKA DISINFORMASI MENYEBAR LUAS, NARASI
KEBENARAN MENJADI TERSINGKIR DAN
TERTUTUPI, DAN PADA AKHIRNYA PUBLIK SULIT
MENEMUKANNYA
DISINFORMASI YANG DILAKUKAN TERUS-MENERUS
OLEH SEBUAH KELOMPOK, YANG MENYENTUH
ASPEK SPIKOLOGIS DAN “POST-TRUTH” PUBLIK,
AKHIRNYA AKAN MENGANCAM DEMOKRASI ”AKAL
PIKIRAN”
GRAFIK JARINGAN SOSIAL INI MEMPERLIHATKAN KETIKA TAGAR
#MAHASISWAPELAJARANARKIS BERHASIL MENEMBUS NARASI
#GEJAYANMEMANGGIL2, YANG IRONISNYA MAHASISWA TURUT
MENGAMPLIFIKASI TAGAR TERSEBUT

84. CONTOH VIRALITAS: OMNIBUS LAW

85. TREN DAN VOLUME ”OMNIBUS LAW”
85

86. SNA ‘OMNIBUS LAW’
86
MENDUKUNG
MENOLAK

87. SNA OMNIBUS LAW (17:00– 22:00)
87
OPOSISI
MEDIA
AKADEMISI
BEM
LSM
AKTIVIS
K-POPERS

88. AVATAR TOP INFLUENCERS /1
88

89. AVATAR TOP INFLUENCERS /2
89

90. TREN OMNIBUS: KLIMAKS DAN EXIT
90
Omnibus Law (sepi)
#MosiTidakPercaya
#TolakOmnibusLaw
#OmnibusLawBawaBerkah
#OmnibusLawBasmiKorupsi
8 Oktober:
Demo Mahasiswa
dan Buruh
13 Oktober:
Demo PA 212
Petinggi KAMI
ditangkap4 Oktober:
RUU Omnibus
Law dibawa ke
Paripurna
5 Oktober:
UU Omnibus Law
Disahkan 16 Oktober:
Demo
Mahasiswa
15 Oktober:
Konpers Pentinggi
KAMI ditangkap
6 Oktober:
Trending
internasional

91. DARI #GAGALKAN OMNIBUS LAW à
#OMNIBUSLAW BASMI KORUPSI
91

92. SNA OMNIBUS LAW (16 OKTOBER 2020)
92
PRO KONTRA
2020

93. ANALISIS BOT DAN SNA

94. MENGETAHUI APAKAH ADA COMPUTATIONAL
PROPAGANDA
• Computational propaganda adalah penggunaan bot atau
algoritma computer untuk mengirim pesan secara otomatis,
dengan tujuan membangun (atau memanipulasi) opini publik.
• Untuk mengetahui sebuah akun itu bot atau bukan, bisa
menggunakan machine learning, yang mengambil beberapa
informasi tentang profile akun tersebut, jaringan sosialnya, serta
aktivitasnya sebagai feature.
• Dengan informasi bot ini, kita bisa mendapat insight:
• Apakah sebuah narasi itu sengaja dibangun oleh pihak tertentu
menggunakan bantuan bot?
• Berapa persen peranan bot dalam sebuah percakapan dibandingkan
dengan akun manusia asli?
• Bagaimana relasi bot dengan akun asli dalam membangun propaganda?
94

95. COMPUTATIONAL PROPAGANDA
95
Since 2012 until now, we have
seen bots, algorithms and other
forms of automation are used by
political actors in countries around
the world to manipulate public
opinion over major social
networking platforms, such as
Twitter, Facebook, Instagram, and
YouTube.

96. BOT ANALYSIS DI DEA
96

97. METODE
97
Botometer
Profile + 200 twits
Bot scores
https://rapidapi.com/OSoMe/api/botometer
API

98. HOW IT WORKS
• Botometer is a machine learning algorithm trained to classify an
account as bot or human based on tens of thousands of labeled
examples.
• When you check an account, you fetches its public profile and
hundreds of its public tweets and mentions using the Twitter API.
• This data is passed to the Botometer API, which extracts about
1,200 features to characterize the account's profile, friends, social
network structure, temporal activity patterns, language, and
sentiment.
• Finally, the features are used by various machine learning models
to compute the bot scores.
98

99. VISUALISASI HASIL BOT ANALYSIS
99
HUMAN
CYBORG
ROBOT

100. AI UNTUK MENDETEKSI ROBOT DI TWITTER
TiLiK
HUMAN ROBOT
HUMAN ROBOT
HUMAN ROBOT
HUMAN ROBOT

101. PETA PERCAKAPAN “JEJAK KHILAFAH”
101
Pro Pemerintah
Pro Oposisi
Pro Khilafah (HTI)
robot
robot

102. CONTOH AKUN ROBOT
102

103. AKSI ROBOT-ROBOT DI TWITTER

104. NEXT…

105. ANALISIS MEDIA SOSIAL LAINNYA
105
• Analisis Trend (WHEN, HOW)
• Analisis Jaringan Sosial (WHO,
HOW, WHAT)
• Analisis Aktor (WHO)
• Analisis Narasi (WHAT, WHY)
• Analisis Topik (WHAT)
• Analisis Geo Location (WHERE)
• Analisis Sentimen (HOW)
• Analisis Emosi (HOW, WHY) https://www.slideshare.net/IsmailFahmi3/social-
media-analytics-dengan-drone-emprit

106. SITASI DRONE EMPRIT

107. HOW TO CITE DRONE EMPRIT?
For Drone Emprit Academic
If you use data directly from Drone Emprit Academic dashboard
(academic.droneemprit.id), use this citation:
Fahmi, I. (2018). Drone Emprit Academic: Software for social media monitoring and
analytics. Available at http://dea.uii.ac.id.
For Drone Emprit
If you use data from Ismail Fahmi's analyses shared on the Internet
(Twitter, Facebook, or Slideshare), use this citation:
Fahmi, I. (2016). Drone Emprit: Software for media monitoring and analytics.
Available at http://pers.droneemprit.id.
107
Source:
https://pers.droneemprit.id/how-to-cite-drone-emprit/

108. DRONE EMPRIT CITATION
108

109. THANK YOU
Ismail Fahmi, PhD.