Social
media
platforms
have
revolutionized
how
people
connect
and
share
information,
but
they
also
bring
challenges
when
it
comes
to
information
dissemination.
Many
studies
in
the
literature
examine
this
phenomenon
using
network
models.
However,
often
focus
on
a
unidimensional
analysis,
considering
only
volume
of
interactions
edges,
which
does
not
fully
capture
different
aspects
phenomenon,
especially
concerning
speed
In
work,
we
propose
framework
that
allows
for
bidimensional
analysis
dissemination
social
platforms,
taking
into
account
both
interactions.
Our
is
based
backbone
extraction
techniques
identify
most
salient
edges
dimensions
classifies
profiles,
allowing
detailed
topology
community
presence
each
profile.
We
applied
two
case
covering
critical
scenarios,
notably
Twitter/X
Telegram.
results
show
proposed
able
uncover
patterns
This
emphasizes
importance
multiple
simultaneously
deeper
understanding
phenomenon.
Proceedings of the International AAAI Conference on Web and Social Media,
Journal Year:
2024,
Volume and Issue:
18, P. 1557 - 1570
Published: May 28, 2024
The
aftermath
of
the
2020
US
Presidential
Election
witnessed
an
unprecedented
attack
on
democratic
values
country
through
violent
insurrection
at
Capitol
Hill
January
6th,
2021.
was
fueled
by
proliferation
conspiracy
theories
and
misleading
claims
about
integrity
election
pushed
political
elites
fringe
communities
social
media.
In
this
study,
we
explore
evolution
content
Twitter
in
seven
months
leading
up
to
attack.
We
examine
suspicious
coordinated
activity
carried
out
users
sharing
content,
finding
evidence
common
adversarial
manipulation
techniques
ranging
from
targeted
amplification
manufactured
consensus.
Further,
map
temporal
of,
relationship
between,
theories,
which
eventually
coalesced
into
rhetoric
a
stolen
election,
with
hashtag
#stopthesteal,
alongside
QAnon-related
narratives.
Our
findings
further
highlight
how
media
platforms
offer
fertile
ground
for
widespread
conspiracies
during
major
societal
events,
can
potentially
lead
offline
actions
organized
violence.
IEEE Access,
Journal Year:
2024,
Volume and Issue:
12, P. 61568 - 61585
Published: Jan. 1, 2024
Online
information
operations
(IOs)
refer
to
organized
attempts
tamper
with
the
regular
flow
of
and
influence
public
opinion.
Coordinated
online
behavior
is
a
tactic
frequently
used
by
IO
perpetrators
boost
spread
outreach
their
messages.
However,
exploitation
coordinated
within
large-scale
IOs
still
largely
unexplored.
Here,
we
build
novel
dataset
comprising
around
624K
users
and4M
tweets
study
howonline
coordinationwas
in
two
recent
carried
out
on
Twitter.
We
investigate
interplay
between
state-of-the-art
network
science
coordination
detection
methods,
providing
evidence
that
both
were
indeed
strongly
coordinated.
Furthermore,
propose
quantitative
indicators
analyses
different
patterns
coordination,
uncovering
malicious
group
managed
hold
central
position
discussion
network,
others
who
remained
at
periphery
limited
interactions
genuine
users.
The
nuanced
results
enabled
our
analysis
provide
insights
into
strategies,
development,
effectiveness
IOs.
Overall,
demonstrate
can
contribute
safeguarding
integrity
platforms.
Scientific Reports,
Journal Year:
2025,
Volume and Issue:
15(1)
Published: Feb. 10, 2025
The
emergence
of
new
public
forums
in
the
form
online
social
media
has
introduced
unprecedented
challenges
to
discourse,
including
polarization,
misinformation,
and
rise
echo
chambers.
Existing
research
extensively
examined
these
topics
by
focusing
on
active
actions
performed
users,
without
accounting
for
share
individuals
who
consume
content
actively
interacting
with
it.
In
contrast,
this
study
incorporates
passive
consumption
data
investigate
prevalence
participation
discourse.
We
introduce
a
metric
quantify
engagement
analyze
over
17
million
pieces
linked
polarized
Twitter
debate
understand
its
relationship
several
features
environments,
such
as
chambers,
coordinated
behavior,
political
bias,
source
reliability.
Our
findings
reveal
significant
proportion
users
interactions,
underscoring
importance
considering
also
proxies
analysis
debates.
Furthermore,
we
found
that
increased
is
primarily
correlated
presence
multimedia
unreliable
news
sources,
rather
than
ideological
stance
producer,
suggesting
independent
work
highlights
significance
quantifying
engagement,
which
influences
platform
feed
algorithms
and,
consequently,
development
discussions.
Moreover,
it
factors
may
encourage
participation,
can
be
utilized
design
more
effective
communication
campaigns.
EPJ Data Science,
Journal Year:
2024,
Volume and Issue:
13(1)
Published: Oct. 7, 2024
Abstract
Automated
social
media
accounts,
known
as
bots,
are
increasingly
recognized
key
tools
for
manipulative
online
activities.
These
activities
can
stem
from
coordination
among
several
accounts
and
these
automated
campaigns
manipulate
network
structure
by
following
other
amplifying
their
content,
posting
messages
to
spam
discourse.
In
this
study,
we
present
a
novel
unsupervised
detection
method
designed
target
specific
category
of
malicious
user
metrics
such
popularity.
Our
framework
identifies
anomalous
patterns
all
the
followers
account.
Through
analysis
large
number
on
Twitter
platform
(rebranded
X
after
acquisition
Elon
Musk),
demonstrated
that
irregular
prevalent
indicative
fake
accounts.
Notably,
found
detected
groups
exhibited
consistent
behavior
across
multiple
This
observation,
combined
with
computational
efficiency
our
proposed
approach,
makes
it
valuable
tool
investigating
large-scale
coordinated
manipulation
platforms.
Scientific Reports,
Journal Year:
2025,
Volume and Issue:
15(1)
Published: May 5, 2025
Malicious
actors
regularly
attempt
to
manipulate
social
media
using
coordinated
posting.
Many
existing
methods
for
detecting
this
coordination,
though,
have
relied
primarily
on
post-timing,
which
is
trivially
easy
change.
In
paper,
we
make
a
significant
methodological
advancement
in
coordination
detection,
leveraging
highly
regular
statistical
patterns
the
speed
and
frequency
of
sharing.
We
apply
validate
approach
Facebook,
11.2
million
link
posts
from
list
16,169
most
popular
English-language
Facebook
pages
that
referenced
at
least
one
top
eight
US
politicians
any
post,
set
produced
more
than
91%
all
user
engagement
category
during
our
collection
period.
Our
can
be
calibrated
adapted
across
contexts,
platforms,
times,
allowing
researchers
build
valid,
testable,
but
still
human-interpretable
models
platform
manipulations.
