The document summarizes research on profiling Bolsobots, which are social media accounts that support Jair Bolsonaro, on Instagram. The researchers mapped the following networks of Bolsobot accounts to analyze their behaviors. Key findings include discrete bots playing a role in content production and spread while avoiding attention, and Bolsobot farms being diverse and adapted to Brazilian social contexts. Profiling bot ecologies can help understand their political influence in Brazil.
Networks, Hashtags, Memes: A Quali-Quantitative Approach for Exploring Social...Janna Joceli Omena
Workshop at CAIS (Center of Advanced Internet Studies), in Bochum. 24 July 2019.
Part 1: Studying Hashtag Engagement through
Digital Networks (and Methods!)
Janna Joceli Omena
Part 2: Situating Internet Memes as Mediators &
Techno-Social Multiplicities
Elena Pilipets
Digital Culture Meets Data: Critical Approaches
˚ ECREA Digital Culture and Communication Section Conference ˚
˚ 6 - 7 November 2017 ˚ University of Brighton ˚ Mithras House ˚
This was the official presentation of SMART; a research group of iNOVA Media Lab. Presentation was held at Universidade Nova de Lisboa (FCSH), on 6 October 2016.
on the ontological necessity of the multidisciplinary development of the webFabien Gandon
Talk on the ontological necessity of the multidisciplinary development of the web at the panel CLOSER/WEBIST 2014 on "social, political and economic implications of cloud and web"
Networks, Hashtags, Memes: A Quali-Quantitative Approach for Exploring Social...Janna Joceli Omena
Workshop at CAIS (Center of Advanced Internet Studies), in Bochum. 24 July 2019.
Part 1: Studying Hashtag Engagement through
Digital Networks (and Methods!)
Janna Joceli Omena
Part 2: Situating Internet Memes as Mediators &
Techno-Social Multiplicities
Elena Pilipets
Digital Culture Meets Data: Critical Approaches
˚ ECREA Digital Culture and Communication Section Conference ˚
˚ 6 - 7 November 2017 ˚ University of Brighton ˚ Mithras House ˚
This was the official presentation of SMART; a research group of iNOVA Media Lab. Presentation was held at Universidade Nova de Lisboa (FCSH), on 6 October 2016.
on the ontological necessity of the multidisciplinary development of the webFabien Gandon
Talk on the ontological necessity of the multidisciplinary development of the web at the panel CLOSER/WEBIST 2014 on "social, political and economic implications of cloud and web"
One Web of pages, One Web of peoples, One Web of Services, One Web of Data, O...Fabien Gandon
Keynote Fabien GANDON, at WIM2016: One Web of pages, One Web of peoples, One Web of Services, One Web of Data, One Web of Things…and with the Semantic Web bind them.
Digital Media Winter Institute 2019
Smart Data Sprint: Beyond visible engagement, Jan. 28 - Feb.1, Universidade Nova de Lisboa, Lisbon, Portugal.
[Short talk]
A glimpse into what social media is all about and how the researchers in the world are using social media. Social media is not a mere hype and not a platform to leverage word-of-the-mouth practices as is the common perception of it in Pakistan: it is much more than that and this is what this talk presented.
Social media is now the place where people are gathering en masse to discuss the news with their friends, neighbors and complete strangers. This change in news consumers’ behavior is proving to be a challenge for local news, but it is also an opportunity. Users and system generated data from social media can also be a boon for content creators. This presentation will feature a case study showing how publishers can use social media analytics to gain insights into their audience and how to use this information to foster a stronger sense of community around their brand of journalism. The case study will focus on how to use Netlytic, a cloud-based social media analytics tool, to mine the public Facebook interactions of the readers of BlogTO, a regional, Canadian-based media outlet, to find out what their readers are interested in and what engages them.
Data Cleaning for social media knowledge extractionMarco Brambilla
Social media platforms let users share their opinions through textual or multimedia content. In many settings, this becomes a valuable source of knowledge that can be exploited for specific business objectives. Brands and companies often ask to monitor social media as sources for understanding the stance, opinion, and sentiment of their customers, audience and potential audience. This is crucial for them because it let them understand the trends and future commercial and marketing opportunities.
However, all this relies on a solid and reliable data collection phase, that grants that all the analyses, extractions and predictions are applied on clean, solid and focused data. Indeed, the typical topic-based collection of social media content performed through keyword-based search typically entails very noisy results.
We recently implemented a simple study aiming at cleaning the data collected from social content, within specific domains or related to given topics of interest. We propose a basic method for data cleaning and removal of off-topic content based on supervised machine learning techniques, i.e. classification, over data collected from social media platforms based on keywords regarding a specific topic. We define a general method for this and then we validate it through an experiment of data extraction from Twitter, with respect to a set of famous cultural institutions in Italy, including theaters, museums, and other venues.
For this case, we collaborated with domain experts to label the dataset, and then we evaluated and compared the performance of classifiers that are trained with different feature extraction strategies.
Fuzzy AndANN Based Mining Approach Testing For Social Network AnalysisIJERA Editor
Fast and Appropriate Social Network Analysis (SNA) tools ,techniques, are required to collect and classify
opinion scores on social networksites , as a grouping on wrong opinion may create problems for a society or
country . Social Network Analysis (SNA) is popular means for researcher as the number of users and groups
increasing day by day on that social sites , and a large group may influence other.In this paper, we
recommendhybrid model of opinion recommendation systems, for single user and for collective community
respectively, formed on social liking and influence network theory. By collecting thedata of user social networks
and preferenceslike, we designed aimproved hybrid prototype to imitate the social influence by like and sharing
the information among groups.The significance of this paper to analyze the suitability of ANN and Fuzzy sets
method in a hybrid manner for social web sites classifications, First, we intend to use Artificial Neural
Network(ANN)techniques in social media data classification by using some contemporary methods different
than the conventional methods of statistics and data analysis, in next we want to propagate the fuzzy approach
as a way to overcome the uncertainity that is always present in social media analysis . We give a brief overview
of the main ideas and recent results of social networks analysis , and we point to relationships between the two
social network analysis and classification approaches .This researchsuggests a hybrid classification model build
on fuzzy and artificial neural network (HFANN). Information Gain and three popular social sites are used to
collect information depicting features that are then used to train and test the proposed methods . This neoteric
approach combines the advantages of ANN and Fuzzy sets in classification accuracy with utilizing social data
and knowledge base available in the hate lexicons.
Social Media Mining - Chapter 10 (Behavior Analytics)SocialMediaMining
R. Zafarani, M. A. Abbasi, and H. Liu, Social Media Mining: An Introduction, Cambridge University Press, 2014.
Free book and slides at http://socialmediamining.info/
Understanding Public Sentiment: Conducting a Related-Tags Content Network Ext...Shalin Hai-Jew
This presentation focuses on how to understand public sentiment through a related-tags content network analysis of public Flickr photos and videos. NodeXL is used to conduct data extractions and visualizations of user-tagged Flickr contents and the resulting “noisy” folksonomies. What mental connections may be made about particular issues based on analysis of text-annotated graphs?
Panel presented as part of the 2017 Data Power Conference (Ottawa, ON, June 23, 2017)
Anatoliy Gruzd (@gruzd), Jenna Jacobson (@jacobsonjenna), Priya Kumar (@link_priya), Philip Mai (@phmai)
Social Media Mining - Chapter 8 (Influence and Homophily)SocialMediaMining
R. Zafarani, M. A. Abbasi, and H. Liu, Social Media Mining: An Introduction, Cambridge University Press, 2014.
Free book and slides at http://socialmediamining.info/
Mining and Comparing Engagement Dynamics Across Multiple Social Media Platfor...The Open University
Understanding what attracts users to engage with social media content is important in domains such as market analytics, advertising, and community management.
To date, many pieces of work have examined engagement dynamics in isolated platforms with little consideration or assessment of how these dynamics might vary between disparate social media systems. Additionally, such explorations have often used different features and notions of engagement, thus rendering the cross-platform comparison of engagement dynamics limited. In this paper we define a common framework of engagement analysis and examine and compare engagement dynamics across five social media platforms: Facebook, Twitter, Boards.ie, Stack Overflow and the SAP Community Network. We define a variety of common features (social and content) to capture the dynamics that correlate with engagement in multiple social media platforms, and present an evaluation pipeline intended to enable cross-platform comparison. Our comparison results demonstrate the varying factors at play in different platforms, while also exposing several similarities.
R. Zafarani, M. A. Abbasi, and H. Liu, Social Media Mining: An Introduction, Cambridge University Press, 2014.
Free book and slides at http://socialmediamining.info/
One Web of pages, One Web of peoples, One Web of Services, One Web of Data, O...Fabien Gandon
Keynote Fabien GANDON, at WIM2016: One Web of pages, One Web of peoples, One Web of Services, One Web of Data, One Web of Things…and with the Semantic Web bind them.
Digital Media Winter Institute 2019
Smart Data Sprint: Beyond visible engagement, Jan. 28 - Feb.1, Universidade Nova de Lisboa, Lisbon, Portugal.
[Short talk]
A glimpse into what social media is all about and how the researchers in the world are using social media. Social media is not a mere hype and not a platform to leverage word-of-the-mouth practices as is the common perception of it in Pakistan: it is much more than that and this is what this talk presented.
Social media is now the place where people are gathering en masse to discuss the news with their friends, neighbors and complete strangers. This change in news consumers’ behavior is proving to be a challenge for local news, but it is also an opportunity. Users and system generated data from social media can also be a boon for content creators. This presentation will feature a case study showing how publishers can use social media analytics to gain insights into their audience and how to use this information to foster a stronger sense of community around their brand of journalism. The case study will focus on how to use Netlytic, a cloud-based social media analytics tool, to mine the public Facebook interactions of the readers of BlogTO, a regional, Canadian-based media outlet, to find out what their readers are interested in and what engages them.
Data Cleaning for social media knowledge extractionMarco Brambilla
Social media platforms let users share their opinions through textual or multimedia content. In many settings, this becomes a valuable source of knowledge that can be exploited for specific business objectives. Brands and companies often ask to monitor social media as sources for understanding the stance, opinion, and sentiment of their customers, audience and potential audience. This is crucial for them because it let them understand the trends and future commercial and marketing opportunities.
However, all this relies on a solid and reliable data collection phase, that grants that all the analyses, extractions and predictions are applied on clean, solid and focused data. Indeed, the typical topic-based collection of social media content performed through keyword-based search typically entails very noisy results.
We recently implemented a simple study aiming at cleaning the data collected from social content, within specific domains or related to given topics of interest. We propose a basic method for data cleaning and removal of off-topic content based on supervised machine learning techniques, i.e. classification, over data collected from social media platforms based on keywords regarding a specific topic. We define a general method for this and then we validate it through an experiment of data extraction from Twitter, with respect to a set of famous cultural institutions in Italy, including theaters, museums, and other venues.
For this case, we collaborated with domain experts to label the dataset, and then we evaluated and compared the performance of classifiers that are trained with different feature extraction strategies.
Fuzzy AndANN Based Mining Approach Testing For Social Network AnalysisIJERA Editor
Fast and Appropriate Social Network Analysis (SNA) tools ,techniques, are required to collect and classify
opinion scores on social networksites , as a grouping on wrong opinion may create problems for a society or
country . Social Network Analysis (SNA) is popular means for researcher as the number of users and groups
increasing day by day on that social sites , and a large group may influence other.In this paper, we
recommendhybrid model of opinion recommendation systems, for single user and for collective community
respectively, formed on social liking and influence network theory. By collecting thedata of user social networks
and preferenceslike, we designed aimproved hybrid prototype to imitate the social influence by like and sharing
the information among groups.The significance of this paper to analyze the suitability of ANN and Fuzzy sets
method in a hybrid manner for social web sites classifications, First, we intend to use Artificial Neural
Network(ANN)techniques in social media data classification by using some contemporary methods different
than the conventional methods of statistics and data analysis, in next we want to propagate the fuzzy approach
as a way to overcome the uncertainity that is always present in social media analysis . We give a brief overview
of the main ideas and recent results of social networks analysis , and we point to relationships between the two
social network analysis and classification approaches .This researchsuggests a hybrid classification model build
on fuzzy and artificial neural network (HFANN). Information Gain and three popular social sites are used to
collect information depicting features that are then used to train and test the proposed methods . This neoteric
approach combines the advantages of ANN and Fuzzy sets in classification accuracy with utilizing social data
and knowledge base available in the hate lexicons.
Social Media Mining - Chapter 10 (Behavior Analytics)SocialMediaMining
R. Zafarani, M. A. Abbasi, and H. Liu, Social Media Mining: An Introduction, Cambridge University Press, 2014.
Free book and slides at http://socialmediamining.info/
Understanding Public Sentiment: Conducting a Related-Tags Content Network Ext...Shalin Hai-Jew
This presentation focuses on how to understand public sentiment through a related-tags content network analysis of public Flickr photos and videos. NodeXL is used to conduct data extractions and visualizations of user-tagged Flickr contents and the resulting “noisy” folksonomies. What mental connections may be made about particular issues based on analysis of text-annotated graphs?
Panel presented as part of the 2017 Data Power Conference (Ottawa, ON, June 23, 2017)
Anatoliy Gruzd (@gruzd), Jenna Jacobson (@jacobsonjenna), Priya Kumar (@link_priya), Philip Mai (@phmai)
Social Media Mining - Chapter 8 (Influence and Homophily)SocialMediaMining
R. Zafarani, M. A. Abbasi, and H. Liu, Social Media Mining: An Introduction, Cambridge University Press, 2014.
Free book and slides at http://socialmediamining.info/
Mining and Comparing Engagement Dynamics Across Multiple Social Media Platfor...The Open University
Understanding what attracts users to engage with social media content is important in domains such as market analytics, advertising, and community management.
To date, many pieces of work have examined engagement dynamics in isolated platforms with little consideration or assessment of how these dynamics might vary between disparate social media systems. Additionally, such explorations have often used different features and notions of engagement, thus rendering the cross-platform comparison of engagement dynamics limited. In this paper we define a common framework of engagement analysis and examine and compare engagement dynamics across five social media platforms: Facebook, Twitter, Boards.ie, Stack Overflow and the SAP Community Network. We define a variety of common features (social and content) to capture the dynamics that correlate with engagement in multiple social media platforms, and present an evaluation pipeline intended to enable cross-platform comparison. Our comparison results demonstrate the varying factors at play in different platforms, while also exposing several similarities.
R. Zafarani, M. A. Abbasi, and H. Liu, Social Media Mining: An Introduction, Cambridge University Press, 2014.
Free book and slides at http://socialmediamining.info/
5 Steps to Marketing Library Services Using Social MediaKatelyn Patterson
Libraries must have a presence on social media in today's world. I presented this at the Texas Library Association's 2015 Annual Conference. Here I will give you 5 steps to marketing your library services using social media.
An overview of how nonprofits are using social media on the web and how others can improve their outreach efforts in a web 2.0 world. It's a "101" program, so it focuses on the basics of networks like Facebook, LinkedIn, blogs, Twitter and Flickr.
A guide to understanding Social Media Monitoring in everyday language. Highlighting Why it is important; what should be done with the data; and how to find a tool.
How and Why Associations should participate in social media. How to begin, What to do first. A basic overview of social media networks and use of media to promote classes, provide information, mediums to use and how to garnish member support and participation.
Grow your association with social media
Debbie Kirkland, Realtor
HomeSalesofTallahassee.com
Twitter for Nonprofits - a presentation made at the Colorado Association of Funders' Social Media Day on August 24, 2010. Includes tips, best practices, thoughts about audience targeting, how to measure results and a look at some key tools and applications that can make using Twitter easier/more effective.
Social Media & Metrics (Digital Marketing Today: F17)Julian Gamboa
In Digital Marketing Today, we strive to make students learn the basics of what will be expected of them in a digital marketing internship. Here, we covered the diverse social media platforms available for companies as well as how to measure growth for periods of time.
Analysing image collections with the computer vision network approachJanna Joceli Omena
Images in Social Media Research:
Digital Tools and Methodological Challenges
Online-Workshop, 10th February 2023.
https://tu-dresden.de/gsw/phil/ikm/kuge/forschung/aktuelle-projekte/bildproteste/news/online-workshop-images-in-social-media-research-digital-tools-and-methodological-challenges
DigiMeth festival, Centre of Interdisciplinary Methodologies at the University of Warwick.
December 9, 2022.
https://warwick.ac.uk/fac/cross_fac/cim/events/digi-meth/
Workshop facilitators: Janna Joceli Omena, Beatrice Gobbo
Abstract:
This workshop offers methodological guidance for narrating networks through visual network analysis (VNA) (Venturini et al. 2021) and a technicity perspective to the practice of digital methods (Omena 2021). It is divided into two parts. First, we will introduce what questions we should ask to make sense of network building and the key principles of VNA. Second, students will work on digital and printed recommendation networks aiming at narrating what they see.
Main takeaways
Students will be able to explore and identify the main components of a digital network
Students will reflect on the distinction between what is network exploration (description tasks) and network narration (insights, findings)
Students will develop the ability to tell a story about the topic under investigation and what constitutes the network.
Requirements:
Please bring your own computer and get familiar with
Retina (https://ouestware.gitlab.io/retina/beta/)
An example of a network 👉 link.
Related projects
Venturini, T., Jacomy, M., & Jensen, P. (2021). What do we see when we look at networks: Visual network analysis, relational ambiguity, and force-directed layouts. Big Data & Society, 8(1). https://doi.org/10.1177/20539517211018488
Omena, J.J.(2021). Digital Methods and Technicity-of-the-Mediums. From Regimes of Functioning to Digital Research. [Doctoral Dissertation, Nova University Lisbon]. Repositório da Universidade Nova de Lisboa. http://hdl.handle.net/10362/127961
Venturini, Tommaso & Bounegru, Liliana & Jacomy, Mathieu & Gray, Jonathan. (2017). 11. How to Tell Stories with Networks Exploring the Narrative Affordances of Graphs with the Iliad: Studying Culture through Data. 10.1515/9789048531011-014.
Making methods with vision APIs, online data & network building (lessons lear...Janna Joceli Omena
Research project on building and interpreting computer vision networks with the purpose to develop visual digital methods for social and media research. Project diary: https://thesocialplatforms.wordpress.com/2020/09/10/computer-vision-networks/
Estudos sobre plataformização mediante a três pilares dos métodos digitaisJanna Joceli Omena
Seminário Internacional Ecologia de Mídia no Contexto da Plataformização
23 a 25 de Setembro, 2021 I Evento virtual I Universidade Federal de Ouro Preto, Minas Gerais, Brasil.
Video: https://www.youtube.com/watch?v=ZXunLB5kfwU&t=22s
Info + videos + grupos organizadores do seminário:
https://www.conjor.com.br/seminarioplataforma
https://r-est.fafich.ufmg.br/seminario-ecologia-de-midias-e-plataformizacao/
The Grammars of Social Media: Thinking platform data under the modes of techn...Janna Joceli Omena
Digital Media Winter Institute 2018
Smart Data Sprint: Interpreters of Platform Data, Jan. 29 - Feb.2, Universidade Nova de Lisboa, Lisbon, Portugal.
[Short talk]
...................................
Correction in slide 17: #DilmaRoussef and #MichelTemer co-related tag Network. See the Follow Network here: https://www.slideshare.net/jannajoceli/why-look-at-social-media-apis-81702316
António Granado and I presented our ongoing work about digital methods for institutional communication and science communication at SciCom PT 2017 - Museu da Ciência da Universidade de Coimbra, Portugal. We explored the case of Portuguese Universities on Facebook.
...................................................................................................
Resumo:
As Universidades Portuguesas no Facebook - Análise de Redes e Comunicação de Ciência
António Granado
Janna Joceli C. de Omena
Como é que as universidades portuguesas fazem uso da rede social Facebook? O quê e como comunicam? Como é que as plataformas digitais servem de ferramenta/ponte para a comunicação de ciência? Qual o contributo da análise visual de redes neste processo? Estas foram as questões-chave que nos conduziram ao presente mapeamento das principais universidades portuguesas no Facebook, a partir de um estudo guiado pela perspetiva dos Métodos Digitais (Rogers, 2013) e de Análise Visual de Redes (Venturini et.al., 2015). O nosso objetivo é apresentar os primeiros resultados obtidos, com base na investigação exploratória que se debruçou sobre duas áreas distintas, mas complementares. Em Março de 2017, com a ajuda da aplicação Netvizz, foram coletados dados das 15 universidades representadas no Conselho de Reitores das Universidades Portuguesas (CRUP). Os dados recolhidos incluíram páginas (todos os posts e interações registadas desde a criação de cada página), imagens da linha do tempo (timeline), e pagelike network. A análise visual de redes permite-nos concluir sobre os principais atores da rede e as suas conexões, assim como autoridade e influência. Para além disto, a análise de clusters possibilita perceber os micro-sistemas de interesse de cada uma das universidades. Após esta primeira análise, selecionámos os posts efetuados pelas 15 universidades durante o ano de 2016 (N=7103), no sentido de obter uma primeira impressão sobre o tipo de conteúdo que estas instituições partilham no Facebook. Os posts foram classificados segundo as 11 categorias propostas por Dumouchel (2014) na sua análise das páginas de Facebook das universidades da Florida (EUA). Deste trabalho, retiramos conclusões iniciais sobre frequência, temas e tipos de conteúdos partilhados pelas universidades portuguesas no Facebook.
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Subhajit Sahu
Abstract — Levelwise PageRank is an alternative method of PageRank computation which decomposes the input graph into a directed acyclic block-graph of strongly connected components, and processes them in topological order, one level at a time. This enables calculation for ranks in a distributed fashion without per-iteration communication, unlike the standard method where all vertices are processed in each iteration. It however comes with a precondition of the absence of dead ends in the input graph. Here, the native non-distributed performance of Levelwise PageRank was compared against Monolithic PageRank on a CPU as well as a GPU. To ensure a fair comparison, Monolithic PageRank was also performed on a graph where vertices were split by components. Results indicate that Levelwise PageRank is about as fast as Monolithic PageRank on the CPU, but quite a bit slower on the GPU. Slowdown on the GPU is likely caused by a large submission of small workloads, and expected to be non-issue when the computation is performed on massive graphs.
As Europe's leading economic powerhouse and the fourth-largest hashtag#economy globally, Germany stands at the forefront of innovation and industrial might. Renowned for its precision engineering and high-tech sectors, Germany's economic structure is heavily supported by a robust service industry, accounting for approximately 68% of its GDP. This economic clout and strategic geopolitical stance position Germany as a focal point in the global cyber threat landscape.
In the face of escalating global tensions, particularly those emanating from geopolitical disputes with nations like hashtag#Russia and hashtag#China, hashtag#Germany has witnessed a significant uptick in targeted cyber operations. Our analysis indicates a marked increase in hashtag#cyberattack sophistication aimed at critical infrastructure and key industrial sectors. These attacks range from ransomware campaigns to hashtag#AdvancedPersistentThreats (hashtag#APTs), threatening national security and business integrity.
🔑 Key findings include:
🔍 Increased frequency and complexity of cyber threats.
🔍 Escalation of state-sponsored and criminally motivated cyber operations.
🔍 Active dark web exchanges of malicious tools and tactics.
Our comprehensive report delves into these challenges, using a blend of open-source and proprietary data collection techniques. By monitoring activity on critical networks and analyzing attack patterns, our team provides a detailed overview of the threats facing German entities.
This report aims to equip stakeholders across public and private sectors with the knowledge to enhance their defensive strategies, reduce exposure to cyber risks, and reinforce Germany's resilience against cyber threats.
Techniques to optimize the pagerank algorithm usually fall in two categories. One is to try reducing the work per iteration, and the other is to try reducing the number of iterations. These goals are often at odds with one another. Skipping computation on vertices which have already converged has the potential to save iteration time. Skipping in-identical vertices, with the same in-links, helps reduce duplicate computations and thus could help reduce iteration time. Road networks often have chains which can be short-circuited before pagerank computation to improve performance. Final ranks of chain nodes can be easily calculated. This could reduce both the iteration time, and the number of iterations. If a graph has no dangling nodes, pagerank of each strongly connected component can be computed in topological order. This could help reduce the iteration time, no. of iterations, and also enable multi-iteration concurrency in pagerank computation. The combination of all of the above methods is the STICD algorithm. [sticd] For dynamic graphs, unchanged components whose ranks are unaffected can be skipped altogether.
Explore our comprehensive data analysis project presentation on predicting product ad campaign performance. Learn how data-driven insights can optimize your marketing strategies and enhance campaign effectiveness. Perfect for professionals and students looking to understand the power of data analysis in advertising. for more details visit: https://bostoninstituteofanalytics.org/data-science-and-artificial-intelligence/
Adjusting primitives for graph : SHORT REPORT / NOTESSubhajit Sahu
Graph algorithms, like PageRank Compressed Sparse Row (CSR) is an adjacency-list based graph representation that is
Multiply with different modes (map)
1. Performance of sequential execution based vs OpenMP based vector multiply.
2. Comparing various launch configs for CUDA based vector multiply.
Sum with different storage types (reduce)
1. Performance of vector element sum using float vs bfloat16 as the storage type.
Sum with different modes (reduce)
1. Performance of sequential execution based vs OpenMP based vector element sum.
2. Performance of memcpy vs in-place based CUDA based vector element sum.
3. Comparing various launch configs for CUDA based vector element sum (memcpy).
4. Comparing various launch configs for CUDA based vector element sum (in-place).
Sum with in-place strategies of CUDA mode (reduce)
1. Comparing various launch configs for CUDA based vector element sum (in-place).
1. Repurposing
Instagram grammar
to profile Bolsobots
Networks
Digital Methods Summer School 2021
Janna Joceli Omena Center for Advanced Internet Studies, Germany Thais Lobo King's College
London, UK Giulia Tucci Federal University of Rio de Janeiro, Brazil Francisco W. Kerche
Federal University of Rio de Janeiro, Brazil Malu Paschoal University of Trento XIaolu JI
Tsinghua University André Rodarte University of Cambridge; Eleanor Griffiths Goldsmiths
College, University of London Lorenzo Zaffaroni Catholic University of Milan Gaia Amadori
Catholic University of Milan Shilan Huang University of Amsterdam, Andrea Medina Density
Design Lab Antonella Autuori Density Design Lab
Artist:
@raphabaggas
2. #Instabots facts
● Social bots are powerful tools that support large-scale advertising strategies, subtly
manipulate public opinion, direct or redirect attention, generate value etc (see also Santini,
Salles, Tucci, 2021);
● Bots networks can be traced by mapping who they are following (DMI, 2020);
● Bots are part of a diverse ecology. They can operate as:
○ Discrete bots: look like real people; exist to like, comment and follow (&
unfollow) other accounts;
○ Ghost accounts: always present, also taken as bridging nodes (Omena 2019, DMI
2020);
○ Private accounts: serve to bot discrete creation (DMI, 2019)
● Tracing following networks through Bolsonaro's non-official accounts can provide
insights on different actions and behaviours that generate dubious content (Omena,
2019).
4. Research objectives
● Map Bolsonaro (non)supportive accounts by following Instagram
recommendations.
● Build and analyse anti/pro-Bolsonaro following network.
● Characterise the collection of botted accounts using i) accounts profile
metadata, ii) following networks, and iii) grouping profile images by colour
similarity.
5. Research questions
● What can the analysis of bolsobots following networks
tell about the agency of political bots in Brazil?
6. Dataset building
● Query design based on Instagram
recommendations of accounts that support
Jair Bolsonaro government;
● Query Instagram search engine for keywords
such as “bolso” or “bolsonaro” or "mito", using
a desktop computer and mobile phone;
● Select profiles by checking a high number of
followers and/or publications;
● Look into selected profiles for
recommendations to other similar accounts;
● Scrape following accounts of seed profiles
7. Dataset building
35 botted accounts
33,771 accounts & image urls
(followed by our seeds)
26,900 unique users
CSV & JSON files:
Profile info metadata
Following accounts list metadata
image urls
8. List of analysis
1. Key grammars: an overview
a. Emojis, URLs, hashtags, geographical distribution, account type;
b. Close reading;
c. Content analysis;
2. Issues alignments & behaviours: in-depth analysis
a. Visual content analysis (image grouping by colour & repetition; qualitative approach);
b. Textual analysis (machine learning & VNA & computer vision networks);
c. Visual Network Analyses (following accounts: monopartite & bipartite graphs)
9. Key Grammars in profile description
What can Instagram features tell about bot ecologies?
10. Bio information: Emojis situated meaning
● The analysis of emojis
appearing more than 20 times
in dataset show that there are
9 main contexts of use:
○ Bolsonaro: support of the
government, patriotic slogans
and eventual campaign in 2022;
○ Family & faith: celebration of
religion and family values;
○ Digital influence: call to action
and community building;
○ Location;
○ Professional;
○ Sports;
○ Agrobusiness;
○ Security;
○ Various.
11. Bio information: URLs
● The high presence of
messaging apps may indicate
that the accounts are using
external platforms to
coordinate action on
Instagram;
● There are civil society political
movements like Nas Ruas
and Vem Pro Quartel;
● And also a religious
evangelical institution, Bola
de Neve.
12. Bio information: Business categories
● Creators & celebrities
are the most common
business categories;
● Followed by
Merchandising stores;
● Religious organizations
and Publishers.
13. Bio information: text word cloud
● Bolsonaro's slogan appear
in the center of the word
cloud;
● Brasil acima de todos Deus
acima de todos / Brazil
above all / God above all
14. Bio information: hashtags
● The high frequency of
hashtags related to
campaign reveal an
already heated debate
around the 2022
presidential election;
● Old campaign's
hashtags (2018):
Instabots need to
update their bio info!
15. Bio information: mentions
● Bolsonaro family as the
most mentioned
profiles;
● Alt-right politicians'
● Conservative media
vehicles;
● Strong support/
recommendations to
Bolsonaro's political
allies.
16. Rio Grande do Sul
Paraná
Santa Catarina
São Paulo
Minas Gerais
Rio de Janeiro
DF
Pernambuco
Ceará
Bio information: Geographical distribution
● Information collected through
the registered mobile phones;
● The geographical distribution
doesn't match the Brazilian
demographics;
● Most accounts are located in
the South region of Brazil,
especially on Rio Grande do Sul
state where there is a tradition
of a more conservative vote
17. Issues alignments & Behaviours
Which themes emerge by clustering bot accounts?
What are the strategies of the bot ecologies?
18. 6 most recurrent words for each topic
LDA on 3.364 unique profiles bios
Bio information text topic modeling
20. Recurrent image:
Alliance for Brasil (Bolsonaro’s party)
Ultra conservative cluster: symbols oust people Call to action cluster: faces and real-like profiles
The center has a wider
range of discourses
Bolsonaro
Call to action
Some examples
21. ● In many cases, we weren't
able to find other versions
of the images in the web;
● One of them seemed to
come from a russian
social network;
Images reverse search analysis
22. Where these images comes from?
Using GVA to see
Shutterstock presence
for ‘regular faces’
23. Parameters used:
1. Ratio following/
follows between
0.8 and 1.2 or
above 5
2. Posts count >
2000
3. Family religious
seems to be the
most unusual
behavior
Spotting the bots
25. 33,768 profile pictures of Bolsobots following network
Instagram, July 2021
grouped by ImageSorter
Visual Analysis: colour grouping & image repetition
26. Categorization as a method to identify and profile bots
Kinds of avatars repetition
in ImageSorter
👥 Followed repetition One user followed by many, so the same profile
appears in ImageSorter multiple times.
🔁 Identical usage repetition One image is used by different users, so the same
image appears in ImageSorter multiple times
🎨 Related aesthetic repetition Usually used by organized bots group
✝ Related symbology repetition Different aesthetics but with related symbols/signs
Account type
Humanized Sophisticated Bot-like It’s personalized, it has stories, it’s linked to a kind of
organization;
🤖 Cheap bot-like It might follow a themes pattern, but it has no posts, or
almost none;
🥸 Discrete bots No profile picture, private account.
Capacity
⬆ High capacity
A mixture of humanistic features including but not
limited to: story highlights, generate a lot of content,
more followers than they are following, replying to
comments, etc.
↕ Medium capacity
⬇ Low capacity
Categorization & behaviour profiling
32. Bot stories: Hammer & Sickle cluster
● Meta-cluster theme:
Anti-leftist
● Account types: Humanised
and sophisticated combined
with discrete/ cheap bot like
● Cyber troop capacity:
Medium
● Messaging strategies:
Attacking the opposition or
mounting smear campaigns
● Colour Cluster: Red
● Type of Repetition: Related
aesthetic repetition
Here is a ‘filler’ or ‘discrete’ bot
within the same cluster. Most
likely sold as a combined
package with the high capacity
bots on the left.
Discrete bots can also offer a
different function to humanised
bots by avoiding attention and
contributing to higher overall
numbers.
Higher capacity bots where the structure of
each account is the same with minor
differences. Posts different content,
different bios, that follow the same theme
33. Bot stories: Anti-Feminist cluster
● Meta-cluster theme:
Anti-leftist
● Account types: Humanised
and sophisticated combined
with discrete/ cheap bot like
● Cyber troop capacity:
Medium
● Messaging strategies:
● Attacking the opposition or
mounting smear campaigns
● Colour Cluster: Red
● Type of Repetition: Related
aesthetic repetition
Misuse of child’s Tiktok content on profile
34. Bot stories: Anti-Feminist cluster
● High capacity bot
interacts with users
impersonating a
child with Tiktok
videos;
● Other users
question the
authenticity of the
profile
Is this you in the video?
Yes it’s me.
I thought the
administrator of this
profile was a full
grown man hahaha
[comment underneath]
Anti.feminista_ replies: hahaha
35. Bot stories: ‘Auntie Bots’ cluster (hearts, flowers, religious avatars)
● Meta-cluster theme:
Personas
● Account types: Humanised
and sophisticated
● Cyber troop capacity: High
● Messaging strategies: No
political content or
Pro-government/pro-party
propaganda
● Colour Cluster: Red, Orange
● Type of Repetition: Related
symbology repetition
Impersonating ‘baby boomers’ online
exploits the clunky and ‘offline’ nature
of boomers on social media.
Therefore very difficult to detect if it is
a bot.
37. Seeds >> Bolsobots
29,693 Instagram accounts
as nodes
37,771 edges as the act of
following other accounts
Maximum number of
following per account: 7,500
Bolsobot bipartite following network
July, 2021
38. Who matters to Bolsobots
supportive accounts?
Who are the bridging accounts?
Where are the ghost accounts
located in network?
How mapping connections
helped us to characterise bots?
39. - Jair Bolsonaro
and his family
- Alt-right politicians
- Government ministries
- The Federal Police of
Brazil
Who matters to Bolsobots supportive accounts?
CENTRE OF THE NETWORK
40. BOTS, BOTS & BOTS
🤖 Conservative accounts & Trump accounts
🤖 Conservative accounts & Extreme-right accounts
🤖 Alliance for Brazil accounts
CELEBRITIES & BRAZILIAN TELEVISION NETWORK
✨📺 Direct relation to SBT Network
Who matters to Bolsobots supportive accounts?
MID-ZONES OF THE NETWORK
44. Where are the
ghost accounts
located in the
network?
GHOST ACCOUNTS I 1.79%
Seeds 0.12% // Remaining accounts 98.1%
Pro-Bolsonaro bots following network I July 2021
50. ● Although very diverse, the ecology of bolsobots profiles is situated in specific
contexts pointing "real life" Bolsonaro's supporting publics;
● Discrete bots play an important role in producing content and sustaining its
spread while maintaining themselves out of the radar;
● Bolsobot farms creation/maintenance are diverse (low/high capacity bots) and
well adapted to Brazilian social stereotypes;
● Profiling bot ecologies can help to build awareness of their agency in the
Brazilian context, contributing to scholarly research in related topics;
Main takeaways