A Real-Time Platform to Monitoring Misinformation on Telegram

Ivandro Claudino, Thiago Gadelha, Tiago Vinuto, José Wellington Franco, José Maria Monteiro

and Javam Machado

Universidade Federal do Ceará, Brazil

jose.monteiro@lsbd.ufc.br, javam.machado@lsbd.ufc.br

Keywords:

Misinformation, Monitoring, Telegram, Social Media.

Abstract:

The large-scale dissemination of misinformation through social media has become a critical issue, harming

public health, social stability and democracy. In Brazil, 79.9% of the population uses social networks, and

the Telegram is present in 65% of the country’s smartphones. Due to its popularization, many groups have

used this instant messaging application to spread misinformation, especially as part of articulated political or

ideological campaigns. Telegram provides two essential features that facilitate the spread of misinformation,

public groups and channels. Through these resources, false information can deceive thousands of people in

a short time. In this context, we present MST, a Real-Time platform to ﬁnd, gather, analyze and visualize

misinformation on Telegram. To evaluate the proposed platform, we built a dataset from the Brazilian general

election campaign in 2022, obtained from Telegram public chat groups and channels.

1 INTRODUCTION

In 2022, the proportion of smartphones with Telegram

installed grew in Brazil from 45% to 60%. Currently,

Telegram instant messaging application is present in

65% of Brazilian’s smartphones

. If on one hand, this

platform offers security and privacy to its users, on

other hand it is an environment with weak or no mod-

eration, which has contributed to the spread of mis-

information. Through Telegram, misinformation can

deceive thousands of people in a very short time and

cause signiﬁcant harm to individuals and society. In

this context, misinformation has been used to change

political scenarios, contribute to the spread of dis-

eases, and even cause deaths (Martins et al., 2022;

Silva and Benevenuto, 2021).

Telegram is a quite popular application due to its

versatility and ease of use. It make it possible to in-

stantly share different media types, such as images,

audios, and videos. Besides, it provide two signiﬁ-

cant features: public chat groups and channels. They

are accessible through invitation links and, usually,

they have speciﬁc topics for discussion, such as pol-

itics and health. Telegram allows users to join or

even share their public groups and channels to simul-

taneously connect to hundreds of people at once, and

quickly receive and share content among themselves.

https://www.mobiletime.com.br/dados-de-mercado/

In this way, public groups and channels are very

similar to social networks. They have been used to

spread misinformation, especially as part of articu-

lated political or ideological campaigns. Furthermore,

misinformation spreads faster, deeper, and expansive

than legit information. Further, due to the high vol-

ume of information that we are exposed to, we have a

limited dexterity to distinguish true information from

misinformation (Martins et al., 2022; Martins et al.,

2021a; Silva and Benevenuto, 2021).

In this context, monitoring the content that circu-

lates in Telegram public groups and channels is a fun-

damental task to identify and understand the spread of

misinformation and get insights to address this prob-

lem. However, collecting a database of messages al-

ready in circulation on Telegram is a challenging task.

To ﬁll this gap, we built the MST, a real-time platform

to monitoring the Misinformation Spreading on Tele-

gram. To evaluate the MST platform, we used it to

build a dataset, concerning the Brazilian general elec-

tions campaign in 2022, obtained from public chat

groups and channels on Telegram.

The remainder of this paper is organized as fol-

lows. Section 2 presents the main related work. Sec-

tion 3 describes the MST platform. Section 4 details

a case study performed in order to evaluate the pro-

posed platform. Conclusions and future work are pre-

sented in Section 5.

Claudino, I., Gadelha, T., Vinuto, T., Franco, J., Monteiro, J. and Machado, J.

A Real-Time Platform to Monitoring Misinformation on Telegram.

DOI: 10.5220/0012039100003467

In Proceedings of the 25th International Conference on Enterprise Information Systems (ICEIS 2023) - Volume 1, pages 271-278

ISBN: 978-989-758-648-4; ISSN: 2184-4992

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

271

2 RELATED WORK

Recently, some efforts towards research involving

Telegram instant messaging application have been de-

veloped. A detailed analysis of Iranian users’ behav-

ior in Telegram was presented in (Hashemi and Cha-

hooki, 2019). More than 900,000 Persian channels

and 300,000 Persian supergroups have been discov-

ered, crawled and inspected in this study. Besides, the

authors devised a method to measure group qualities

in Telegram. In (Ng and Loke, 2021), the authors ana-

lyzed a Singapore-based COVID-19 Telegram group

with more than 10,000 participants focusing on ﬁve

dimensions: participation, sentiment, negative emo-

tions, topics, and message types. In (Baumgartner

et al., 2020), the authors collected a large volume of

data composed of almost 28,000 channels on Tele-

gram, giving rise to a useful data set that can be used

to study politics, protests, online social movements

and disinformation in the context of mobile applica-

tions. During the process the authors collected data

and metadata from public channels and from this they

built a list of channels focused on important topics,

such as right-wing extremist politics. In the Baum-

gartner et al.’s work (Baumgartner et al., 2020) both

the data and the source code used to collect it, were

made available.

Paz et al. (de Paz et al., 2022) performed an anal-

ysis about disinformation spreading on Telegram and

proposed a way to identify the origin of the published

content. The work presented in (Nobari et al., 2021)

provided an in-depth look at how messages conveyed

from telegram go viral. The authors performed a

study based on a real dataset obtained from Telegram.

They looked several aspects such as, the information

ﬂow and the characteristics of viral messages. In

(de Paz et al., 2023), the authors presented an anal-

ysis of misinformation cross-platform dynamics by

focusing on communications published by COVID19

negationists on Twitter and Telegram. Herasimenka et

al. (Herasimenka et al., 2022) analyzed 200,000 Tele-

gram posts and observed that links to known sources

of misleading information are shared more often than

links to professional news content, but the former

stays conﬁned to relatively few channels. They also

concluded that, contrary to popular received wisdom,

the audience for misinformation is not a general one,

but a small and active community of users. Akbari et

al. (Akbari and Gabdulhakov, 2019) investigated the

Telegram ban in Russia and Iran, when both govern-

ments demanded access to the content shared by the

users and the platform refused to provide it. Besides,

they provided an overview of the actors, methods, and

tools that are instrumentalized against Telegram.

In (Júnior et al., 2021) the authors collected

1,405,997 messages from 122 Brazilian political

groups and channels on Telegram, covering the pe-

riod from January 2018 to April 2021. Besides, using

this dataset, they performed some data analysis tech-

niques, observing the network created in the platform

as well as a closer look in the dynamics of messages

and members in this platform. Their ﬁndings showed

that the political discussion on Telegram took a leap

in the beginning of 2021 (up to 3 times compared to

the end of 2020 in the number of messages). In addi-

tion, the authors observed a signiﬁcant volume of re-

ferrals on Telegram, as well as a signiﬁcant amount of

links to content on other platforms such as YouTube,

evidencing the use of Telegram as a vector for con-

tent sharing. They concluded that the large groups

structure of Telegram are effective in spreading the

messages through the network, with the content be-

ing viewed by numerous users and forwarded mul-

tiple times. Furthermore, they observed a relevant

amount of messages attacking political personalities

and spreading unchecked content about COVID-19

pandemic.

In (Dargahi Nobari et al., 2017), the authors de-

veloped a crawler to collect public data from Tele-

gram, including messages, users, groups, channels

and their relationships. In addition, they created a

mention graph and applied the page rank algorithm

in order to understand the differences concerning

link patterns between Telegram and other networks.

In (Paschalides et al., 2020), the authors presented

MANDOLA, a big-data processing system that mon-

itors, detects, visualizes, and reports the spread and

penetration of online hate-related speech using big-

data approaches. MANDOLA consists of six com-

ponents that intercommunicate to consume, process,

store, and visualize statistical information regard-

ing hate speech spread online. They also proposed

a novel ensemble-based classiﬁcation algorithm for

hate speech detection. Khaund et al. (Khaund et al.,

2021) presented a methodology to collect and analyze

data from Telegram. They conducted both text and

network analysis to gain insights into political dis-

course and public opinion. Their ﬁndings included

the use of Telegram by politicians to connect with

their voter, promote their work as well as ridicule their

peers. Besides, the channels were actively used to

disseminate information on political affairs while the

chat groups to discuss views about the government.

Benevenuto et al. presented in (Júnior et al., 2022a;

Júnior et al., 2022b) the “Telegram Monitor”, a web-

based system that monitors the political debate in this

platform and enables the analysis of the most shared

content in multiple channels and public groups.

ICEIS 2023 - 25th International Conference on Enterprise Information Systems

272

3 THE MST PLATFORM

The MST (Misinformation Spreading on Telegram)

platform architecture comprises nine components, as

illustrated in Figure 1. Next, we will discuss in detail

each one of these components.

3.1 Telegram Connector

The main goal of this component is make possible

gathering data from Telegram using a common inter-

face. It converts each captured message to the JSON

(JavaScript Object Notation) format and send it to

the Message Broker (Redis). We chose JSON be-

cause it is a language-independent, standard format

for storing and exchanging data. Besides, the Tele-

gram Connector sends each captured media ﬁle to the

File Server component. In order to avoid storing du-

plicate media ﬁles, we apply the MD5 hash algorithm

on the ﬁle content and generate a unique identiﬁer,

which is used as the name of the media ﬁle. Thus, we

avoid wasting disk space, as well as making it pos-

sible to aggregate similar content and quantify how

many times each one was shared by the users, with

the purpose of understand the popularity of each con-

tent. Listing 1 illustrates an example of JSON “ﬁle”

caught the Telegram Connector.

3.1.1 Finding Groups and Channels

The ﬁrst step in collecting data from Telegram is to

ﬁnd groups and channels of interest. We divide this

task into seven steps:

• Step 1: Initially, a manual search is per-

formed on Telegram grouping links of pub-

lic channels and chat groups based on a list

of keywords (as our case study aims to inves-

tigate the Brazilian far-right groups, we used

terms such as: bolsonaro, captain, patriots).

So, based on how Telegram group invitation

links are structured, we searched for URL pat-

terns such as https://t.me/joinchat/<GroupID>

and https://telegram.me/<GroupID> on popular

search engines (e.g. Google), besides we also

looked for the terms in the keywords list using

Telegram’s search feature. Then, we insert the

found links in a ﬁle called “CandidateGroups”;

• Step 2: We performed a search for Telegram invi-

tation links in public datasets of Whatsapp mes-

sages like (Martins et al., 2022; Martins et al.,

2021a; Martins et al., 2021b; de Sá et al., 2021).

Then, we insert the found links in the “Candidate-

Groups” ﬁle;

• Step 3: In this step, from the “CandidateGroups”

ﬁle a manual selection of groups and channels that

are inline with the research goals was performed,

that is, the selected groups should be focused on

the Brazilian far-right themes.

• Step 4: We then joined the selected groups with a

new Telegram account.

• Step 5: We collect the messages that travel in the

selected groups through the Telegram API.

• Step 6: Weekly, we parse the collected messages

searching for new invitation links, that is links

that re are not previously collected, and insert the

found links in a new version of the “Candidate-

Groups” ﬁle.

• Step 7: Finally, we repeat the steps 3 to 6 for the

“CandidateGroups” ﬁle built at the previous step.

3.1.2 Gathering Data

This subsection brieﬂy describes how the MST plat-

form collects the content that travel in groups and

channels using the Telegram ofﬁcial API. So, to col-

lect these contents, we created a Telegram account

for this research and developed an application

writ-

ten in Python using the Telethon library

(library that

aims to facilitate integration with the Telegram API).

This Python applications obtains metadata concerning

groups and/or channels and users, besides text mes-

sages a media ﬁles (images, audios and videos).

The Telethon library allows the creation of a client

to establish a connection between the MST platform

and the Telegram environment. The purpose of this

connection is to keep the client running as long as

the connection is not interrupted. As we aim to cap-

ture text messages and media content in real time, we

chose to keep the client running constantly. An event

handler in the client is triggered when there is a no-

tiﬁcation of a new message. Thus, the client receives

the identiﬁer and the textual content of this new mes-

sage. Next, the client uses Telethon library to ob-

tains more detailed information, including text mes-

sage, date, time, among others. With this informa-

tion, the client builds a JSON ﬁle and send it to the

Message Broker, as illustrated in Listing 1. Despite

Telethon library abstracts and simpliﬁes interactions

with the Telegram API, it uses a speciﬁc class for dif-

ferent type of content (images, audio/video, URLs,

among others) and for each type of chat. In addition,

these different classes have some distinct attributes.

This fact was an important challenge for coding the

Telegram Connector.

https://core.telegram.org/#getting-started

https://github.com/LonamiWebs/Telethon

A Real-Time Platform to Monitoring Misinformation on Telegram

273

Figure 1: MST Architecture.

3.2 Message Broker

Message Broker is a software that make it possible

that applications, systems and services communicate

with each other and share information. It is respon-

sible for validating, storing, routing and delivering

messages to the appropriate destinations. In the MST

platform, the Message Broker acts as an intermediary,

allowing the Telegram Connector to send messages to

the ETL Application. This facilitates the decoupling

of components within the proposed architecture. The

Message Broker allows reliable storage and ensures

message delivery. It has a set of message queues,

which store and sort messages until the ETL Appli-

cation can process them. Furthermore, it ensures that

each queued message is consumed only once. To im-

plement the Message Broker, we use Redis, which is

an in-memory, key-value, open source, versatile and

easy-to-use storage system. In addition, it provides

high performance, persistence and data replication.

3.3 ETL Application

The ETL Application is responsible for text mes-

sages processing, which includes different tasks, such

as parsing, anonymization, misinformation detection

and sentiment analysis. Many of these tasks use the

services of the Data Processing API. We took into

consideration privacy issues by anonymizing users’

names and cell phone numbers. For this, we create

an anonymous and unique ID for each user by using

an MD5 hash function on their username. Similarly,

we create an anonymous alias for each group. Af-

ter a text message processing, the ETL Application

component sends the resulting data to a Relational

Database Server (PostgreSQL) and a Search Engine

(Elasticsearch).

3.4 File Server

The File Server component is responsible for store the

media ﬁles (e.g. audios, images and videos) captured

by the Telegram Connector in a persistent and safe

manner. In order to avoid to store several copies of

the same content, we used the following strategy:

• First, we apply the MD5 algorithm on the bytes of

the media content, which are present in the meta-

data provided by the Telegram API, and generate

a hash code (an unique identiﬁer) that is used as

the name of the media ﬁle on the server;

• Next, we verify if there is a ﬁle with the same

name of the hash code generate previously. In

negative case, we download and store the media

ﬁle. In afﬁrmative case, we don’t download the

media ﬁle, since it is already stored in our server,

avoid processing and storing overhead.

However, this strategy doesn’t work for some con-

tent types, such as SVG images and MKV videos,

since their bytes are not present in the metadata pro-

vided by the Telegram API. In this case, we down-

load the media ﬁle, generate the hash code, and then

we check if this content has been previously down-

loaded. In afﬁrmative case, we delete this media ﬁle.

ICEIS 2023 - 25th International Conference on Enterprise Information Systems

274

{

" id_ me ss ag e ": "104 981 9 " ,

" mes s e n g e r ": " tel eg ra m " ,

" mes sa g e _ ty pe ": " Ur l ",

" id_ pe rs on a ": #### ### # # ,

" dat e_ m e s sa ge ": "202 2 - 09 -28 T16 : 1 8: 09 + 00 :0 0" ,

" tex t_ c o n te nt ": " Banc o M un di a l con fi r ma fala de Pa ulo Gue d es : PIB do Br a si l t e nde

a c re sc er ma is que o da Chin a \ u26 1b h ttp s :// t er r a b ra s il n ot i c i as . c o m /2 02 2/ 09 /

banco - m undial - c onfi r ma - falaa -de - p aulo - guedes - pib -do - br a s il - t ende -a - cr escer - mais

- que - o - da - china /\ n \ n Bom dia \ u d83 d \ udd 2 5 \ ud8 3d \ ud d 25 \n \ n@ F Y IB RA SI L ",

" id _ m em b er _ t e le g ra m ": ### ### # ## ,

" id_g ro up ": ## # ### ### ,

" m edi a ": " c a9 a 6 c 2 5 f 2 9 52 6 93 0 d6 0 8 5 2 a 74 a b6 9 40 (1) . jp g ",

" med ia _n am e ": "" ,

" med ia _t yp e ": " url ",

" med i a _ u r l ": " https :/ / ter r ab r a s il n ot i ci a s . com / 20 22 /0 9/ banco - mund ial - co n firm a -

falaa - de - paulo - guedes - pib - do - bra s i l - t ende -a - cr escer - mais - que -o - da - ch i na /" ,

" med i a _ m d 5 ": " ca 9 a6 c 25 f 2 9 5 2 69 3 0d 6 08 5 2 a 7 4 ab 6 94 0 ",

" dis pl a y _ na me ": " Banc o M un di a l con fi r ma fala de Pa ulo Gue d es : PIB do Br a si l t e nde

a c re sc er ma is que o da Chin a - Te rra Bra s il Not \ u 00 ed ci as \ nCo mp a r t il he : A

nova prev is \ u 00e 3o do Ba nco Mun di a l \ u 00e 9 de que o PIB ( P ro dut o Inte rn o B rut o )

" add r es s_ m es sa g e ": "" ,

" lat i tu d e _ me s s a ge ": 0 ,

" lo n g it u d e _m e ss ag e ": 0 ,

" con t ac t s _ me s s a ge ": nu l l

}

Listing 1: Example of a JSON Caught from Telegram.

In negative case, we don’t need do anything.

3.5 Database

The Relational Database Server supports storing and

querying data on the traditional ﬂat model. There-

unto, it uses PostgreSQL

, a free and open-source re-

lational database management system (RDBMS) em-

phasizing extensibility and SQL compliance. It is

important to highlight that the audios, images, and

videos are stored by the File Server. The PostgreSQL

database stores only the path to these ﬁles.

3.6 Search Engine

The Search Engine component aims to provide tex-

tual queries directly on the captured messages. For

this, it uses Elasticsearch

, a search engine based

on the Lucene library that provides a distributed,

multitenant-capable full-text search engine with an

HTTP web interface and schema-free JSON docu-

ments.

https://www.postgresql.org/

https://www.elastic.co/elasticsearch/

3.7 Web Portal

Today, there is a great need for displaying massive

amounts of data in a way that is easily accessible and

understandable. In this context, data visualization is

a way to represent information graphically, highlight-

ing patterns and trends in data and helping to achieve

news insights. It enables the data exploration via the

manipulation of charts and images. More speciﬁ-

cally, it enables users to analyze the data by interact-

ing directly with a visual representation of it. In this

work, the Web Portal component is a web application

developed using Python programming language and

Django 3 framework, which explores relational (from

PostgreSQL) and textual (from Elasticsearch) data.

3.8 Telegram Bot

Telegram Bot is a proactive chatbot built from Tele-

gram which automatically detects and alerts the pres-

ence of misinformation in social chats. Initially, they

need to be added to a certain group. Then it will auto-

matically monitor and analyze the content that travels

in the group. Finally, if they detect that certain content

has a high probability of containing misinformation,

an alert message is sent to the group.

A Real-Time Platform to Monitoring Misinformation on Telegram

275

4 CASE STUDY

To evaluate the MST platform, we performed an ex-

ploratory case study using a dataset covering the

Brazilian general elections campaign in 2022 col-

lected by Telegram. We used the MST platform to

monitor right-far Telegram groups and channels in the

period between September 27, 2022, and November

15, 2022, which comprises part of the electoral pe-

riod of the 1st round and all of the 2nd round of 2022

general Brazilian elections. This case study was in-

ﬂuenced by (de Sá et al., 2021; Júnior et al., 2022a).

So, many data analysis techniques were applied to this

dataset to get insights about misinformation spread-

ing. The dataset built contains 513,961 messages ob-

tained from 14,085 users from 180 Groups/Channels.

4.1 Messages Characterization

Initially, we will present some visualizations to char-

acterize the built dataset. Figure 2 shows the distri-

bution of messages sending time by the day hours on

Telegram. As we can imagine, the peak of sending

messages occurs at the time reserved for lunch (be-

tween 12 and 15 hours) and in the early evening, just

after work hours. Figure 3 shows the distribution mes-

sages sent time by day on Telegram. As we can imag-

ine, the peak of sending messages occurs on October

2nd (the date of the ﬁrst round of elections) and Oc-

tober 30th (the date of the second round of elections).

Figure 2: Number of Messages by Hour.

Figure 3: Number of Messages by day.

4.2 Vocabulary Characterization

Another aspect that needs to be analyzed is the char-

acteristics of the vocabulary used in the text messages

since there is a strong relationship between the used

vocabulary and the social network, in this case, Tele-

gram. Figure 4 shows the word cloud highlighting the

most popular words on monitored Telegram public

chat groups and channels. Deserve mention the terms

related with the Brazilian far-right, such as: Bol-

sonaro, presidente, Brasil, forças armadas, exército,

direira, verdade, Deus and família. Furthermore, we

can highlight terms related with the far-right oppo-

nents, such as: Lula, esquerda, petista, comunista,

bandido and ladrão. Finally, we can observe terms

related to conspiracy theories, such as: TSE, fraude,

urna, voto and Alexandre de Moraes.

Figure 4: Word Cloud on Telegram.

4.3 Misinformation Analysis

In order to evaluate the misinformation potential in

the used dataset, we applied the misinformation clas-

siﬁer proposed in (Cabral et al., 2021) on the text mes-

sages caught by MST. It is important to highlight that

this classiﬁer was trained with a dataset collected and

labelled during the 2018 Brazilian General Elections.

This misinformation classiﬁer receives a text message

as input and generates as output a value between 0 and

1, indicating the probability that the text contains mis-

information. Texts with values greater than 0.5 were

considered misinformation, messages with values be-

tween 0.3 and 0.5 were designated inconclusive (neu-

tral), while texts with values less than 0.3 were treated

as non-misinformation. Figure 5 shows that the used

classiﬁer considered that 25.31% of the caught text

messages contained misinformation. Figure 5 shows

the distribution of the misinformation probability.

The creators of misinformation use various stylis-

tic tricks to promote the success of their contents,

ICEIS 2023 - 25th International Conference on Enterprise Information Systems

276

Figure 5: Proportion of Misinformation.

Figure 6: Distribution of the Misinformation Probability.

with one of them being to excite the sentiments of the

recipients. Often, misinformation is associated with

the presence polarity (positive and negative), which is

used by content producers to misdirect readers. Con-

sequently, Sentiment Analysis (SA) provides crucial

information on the content of a Telegram message to

determine whether it is trustworthy or should be con-

sidered as misinformation.

In this context, we applied the LeIA tool on the

text messages caught by MST platform. LeIA re-

ceives a text message as input and generate as output

a value between -1 and 1, called sentiment score. Val-

ues close to 1 indicate high positivity, values close

to -1 denote high negativity, and values close to 0

imply neutrality. Figure 7 shows the proportion be-

tween positive, negative and neutral messages. Texts

with sentiment score greater then 0.3 where consid-

ered “Positive”. Messages with sentiment score less

then -0.3 where designated “Negative”. Texts with

sentiment score between -0.3 and 0.3 were treated as

“neutral”. Analyzing Figure 7, we can note that al-

most half of the messages have high polarity (“Pos-

itive” or “Negative”). Messages with high polarity

could be used to generate alerts or should be given

priority when going through veracity veriﬁcation pro-

cesses. Figure 8 illustrates the sentiment score distri-

bution.

Figure 7: Proportion of Sentiment Analysis.

Figure 8: Distribution of the Sentiment Analysis.

5 CONCLUSIONS

Due to its popularization, many groups have used

Telegram to spread false information, especially as

part of articulated political or ideological campaigns.

The large-scale and fast spread of misinformation

through Telegram messages poses a signiﬁcant social

problem, harming public health, social stability and

democracy. Telegram provides two essential features

that facilitate the spread of misinformation: public

groups and channels. Through these resources, misin-

formation can deceive thousands of people in a short

time. In this context, we presented MST (Misinfor-

mation Spreading on Telegram), a Real-Time plat-

form to ﬁnd, gather, analyze and visualize misinfor-

mation on Telegram. To evaluate the proposed plat-

form, we built a dataset from the Brazilian general

election campaign in 2022, obtained from Telegram

public chat groups and channels. The MST platform

and the built dataset are available at our public repos-

itory

. We hope that MST platform can help journal-

ists and researchers to understand the misinformation

propagation in Telegram. As future works, we want

to analyze how misinformation spread among the chat

groups using complex networks.

https://gitlab.com/jmmonteiro

A Real-Time Platform to Monitoring Misinformation on Telegram

277

REFERENCES

Akbari, A. and Gabdulhakov, R. (2019). Platform surveil-

lance and resistance in iran and russia: The case of

telegram. Surveillance & Society, 17:223–231.

Baumgartner, J., Zannettou, S., Squire, M., and Blackburn,

J. (2020). The pushshift telegram dataset. In Proceed-

ings of the International AAAI Conference on Web and

Social Media, volume 14, pages 840–847.

Cabral, L., Monteiro, J. M., da Silva, J. W. F., Mattos, C.

L. C., and Mourão, P. J. C. (2021). Fakewhastapp.br:

NLP and machine learning techniques for misinfor-

mation detection in brazilian portuguese whatsapp

messages. In Filipe, J., Smialek, M., Brodsky, A.,

and Hammoudi, S., editors, Proceedings of the 23rd

International Conference on Enterprise Information

Systems, ICEIS 2021, Online Streaming, April 26-28,

2021, Volume 1, pages 63–74. SCITEPRESS.

Dargahi Nobari, A., Reshadatmand, N., and Neshati, M.

(2017). Analysis of telegram, an instant messaging

service. In Proceedings of the 2017 ACM on Con-

ference on Information and Knowledge Management,

pages 2035–2038.

de Paz, A., Suárez, M., Palmero, S., Degli-Esposti, S., and

Arroyo, D. (2022). Following negationists on twitter

and telegram: Application of ncd to the analysis of

multiplatform misinformation dynamics. In Proceed-

ings of the International Conference on Ubiquitous

Computing & Ambient Intelligence (UCAmI 2022),

pages 1110–1116. Springer.

de Paz, A., Suárez, M., Palmero, S., Degli-Esposti, S., and

Arroyo, D. (2023). Following negationists on twit-

ter and telegram: Application of ncd to the analysis

of multiplatform misinformation dynamics. In Bravo,

J., Ochoa, S., and Favela, J., editors, Proceedings

of the International Conference on Ubiquitous Com-

puting & Ambient Intelligence (UCAmI 2022), pages

1110–1116, Cham. Springer International Publishing.

de Sá, I. C., Monteiro, J. M., da Silva, J. W. F., Medeiros,

L. M., Mourão, P. J. C., and da Cunha, L. C. C.

(2021). Digital lighthouse: A platform for monitor-

ing public groups in whatsapp. In Filipe, J., Smi-

alek, M., Brodsky, A., and Hammoudi, S., editors,

Proceedings of the 23rd International Conference on

Enterprise Information Systems, ICEIS 2021, Online

Streaming, April 26-28, 2021, Volume 1, pages 297–

304. SCITEPRESS.

Hashemi, A. and Chahooki, M. A. Z. (2019). Telegram

group quality measurement by user behavior analysis.

Social Network Analysis and Mining, 9:1–12.

Herasimenka, A., Bright, J., Knuutila, A., and Howard,

P. N. (2022). Misinformation and professional news

on largely unmoderated platforms: the case of tele-

gram. Journal of Information Technology & Politics,

0(0):1–15.

Júnior, M., Melo, P., da Silva, A. P. C., Benevenuto, F., and

Almeida, J. (2021). Towards understanding the use

of telegram by political groups in brazil. In Proceed-

ings of the Brazilian Symposium on Multimedia and

the Web, pages 237–244.

Júnior, M., Melo, P. F., Kansaon, D., Mafra, V., Sá, K.,

and Benevenuto, F. (2022a). Telegram monitor: Mon-

itoring brazilian political groups and channels on tele-

gram. CoRR, abs/2202.04737.

Júnior, M., Melo, P. F., Kansaon, D., Mafra, V., Sá, K., and

Benevenuto, F. (2022b). Telegram monitor: Moni-

toring brazilian political groups and channels on tele-

gram. In Bellogín, A., Boratto, L., and Cena, F., edi-

tors, HT ’22: 33rd ACM Conference on Hypertext and

Social Media, Barcelona, Spain, 28 June 2022- 1 July

2022, pages 228–231. ACM.

Khaund, T., Hussain, M. N., Shaik, M., and Agarwal,

N. (2021). Telegram: Data collection, opportunities

and challenges. In Lossio-Ventura, J. A., Valverde-

Rebaza, J. C., Díaz, E., and Alatrista-Salas, H., ed-

itors, Information Management and Big Data, pages

513–526, Cham. Springer International Publishing.

Martins, A. D. F., Cabral, L., Mourão, P. J. C., Monteiro,

J. M., and Machado, J. (2021a). Detection of misinfor-

mation about covid-19 in brazilian portuguese what-

sapp messages. In International Conference on Appli-

cations of Natural Language to Information Systems,

pages 199–206. Springer.

Martins, A. D. F., da Cunha, L. C. C., Mourão, P. J. C.,

de Sá, I. C., Monteiro, J. M., and de Castro Machado,

J. (2021b). Covid19.br: A dataset of misinformation

about covid-19 in brazilian portuguese whatsapp mes-

sages. In III Dataset Showcase Workshop, DSW 2021,

Rio de Janeiro, RJ, Brazil, October 4-8, 2021 (To ap-

pear). SBC.

Martins, A. D. F., Monteiro, J. M., and Machado, J. C.

(2022). Understanding misinformation about COVID-

19 in whatsapp messages. In Chiusano, S., Cerquitelli,

T., Wrembel, R., Nørvåg, K., Catania, B., Vargas-

Solar, G., and Zumpano, E., editors, New Trends

in Database and Information Systems - ADBIS 2022

Short Papers, Doctoral Consortium and Workshops:

DOING, K-GALS, MADEISD, MegaData, SWODCH,

Turin, Italy, September 5-8, 2022, Proceedings, vol-

ume 1652 of Communications in Computer and Infor-

mation Science, pages 14–23. Springer.

Ng, L. H. X. and Loke, J. Y. (2021). Analyzing public opin-

ion and misinformation in a covid-19 telegram group

chat. IEEE Internet Computing, 25(2):84–91.

Nobari, A. D., Sarraf, M. H. K. M., Neshati, M., and

Daneshvar, F. E. (2021). Characteristics of viral mes-

sages on telegram; the world’s largest hybrid public

and private messenger. Expert systems with applica-

tions, 168:114303.

Paschalides, D., Stephanidis, D., Andreou, A., Orphanou,

K., Pallis, G., Dikaiakos, M. D., and Markatos, E.

(2020). Mandola: A big-data processing and visu-

alization platform for monitoring and detecting online

hate speech. ACM Trans. Internet Technol., 20(2).

Silva, M. and Benevenuto, F. (2021). COVID-19 ads as po-

litical weapon. In Hung, C., Hong, J., Bechini, A., and

Song, E., editors, SAC ’21: The 36th ACM/SIGAPP

Symposium on Applied Computing, Virtual Event, Re-

public of Korea, March 22-26, 2021, pages 1705–

1710. ACM.

ICEIS 2023 - 25th International Conference on Enterprise Information Systems

278