Medfluencer: A Network Representation of Medical Influencers’ Identities and Discourse on Social Media

Zhijin Guo; Edwin D. Simpson; Roberta Bernardi

Medfluencer: A Network Representation of Medical Influencers’ Identities and Discourse on Social Media

Research output: Contribution to conference › Conference Paper › peer-review

192 Downloads (Pure)

Abstract

In our study, we first constructed a dataset from the tweets ofthe top 100 medical influencers with the highest Influencer Score[14] during the COVID-19 pandemic. This dataset was then usedto construct a socio-semantic network, mapping both their identities and key topics, which are crucial for understanding theirimpact on public health discourse. To achieve this, we developeda few-shot multi-label classifier to identify influencers and theirnetwork actors’ identities, employed BERTopic for extracting thematic content, and integrated these components into a networkmodel to analyze their impact on health discourse. To ensure thereproducibility of our results, we have made the code available athttps://github.com/ZhijinGuo/Medinfluencer.

Original language	English
Number of pages	8
Publication status	Published - 4 Jul 2024
Event	epiDAMIK 2024: : The 7th International workshop on Epidemiology meets data Miningand Knowledge discovery - https://epidamik.github.io/, Barcelona, Spain Duration: 26 Aug 2024 → 26 Aug 2024

Workshop

Workshop	epiDAMIK 2024:
Abbreviated title	epiDAMIK
Country/Territory	Spain
City	Barcelona
Period	26/08/24 → 26/08/24

Research Groups and Themes

Jean Golding
social network
Covid-19
large language models
social media
MGMT Operations and Management Science
social network
Covid-19
large language models
social media
Elizabeth Blackwell Institute
social network
Covid-19
large language models
social media
Digital Societies
social network
Covid-19
large language models
social media
MGMT theme Innovation and Digitalisation
social network
Covid-19
large language models
social media
MGMT theme Public Services Governance and Management
social network
Covid-19
large language models
social media
Health and Wellbeing
social network
Covid-19`
large language models
social media

Keywords

social network
Covid-19
large language models (LLMs)
social media

Access to Document

Full-text PDF (accepted manuscript) Accepted author manuscript, 1.29 MBLicence: CC BY

https://openreview.net/pdf?id=StgezOmzHdLicence: CC BY

Handle.net

Persistent link

“Medfluencer: a socio-semantic network analysis method to investigate how medical influencers affect health discourse on social media
Bernardi, R. (Principal Investigator), Simpson, E. D. (Co-Principal Investigator) & Guo, Z. (Researcher)
20/11/23 → 31/07/24
Project: Research

University of Bristol Research Fellowship
Bernardi, R. (Recipient), 2023
Prize: Prizes, Medals, Awards and Grants

Invited Talk on “Who says What? Identities and NLI for Studying Health Misinformation on Social Media”
Bernardi, R. (Participant)
4 Dec 2024
Activity: Participating in or organising an event types › Invited talk
Invited Talk on “Who says What? Identities and NLI for Studying Health Misinformation on Social Media,” WORDS Network Autumn Meeting.
Bernardi, R. (Participant)
26 Nov 2024
Activity: Participating in or organising an event types › Invited talk

Cite this

@conference{8461593e11574bc080c9c462b60f4884,

title = "Medfluencer: A Network Representation of Medical Influencers{\textquoteright} Identities and Discourse on Social Media",

abstract = "In our study, we first constructed a dataset from the tweets ofthe top 100 medical influencers with the highest Influencer Score[14] during the COVID-19 pandemic. This dataset was then usedto construct a socio-semantic network, mapping both their identities and key topics, which are crucial for understanding theirimpact on public health discourse. To achieve this, we developeda few-shot multi-label classifier to identify influencers and theirnetwork actors{\textquoteright} identities, employed BERTopic for extracting thematic content, and integrated these components into a networkmodel to analyze their impact on health discourse. To ensure thereproducibility of our results, we have made the code available athttps://github.com/ZhijinGuo/Medinfluencer.",

keywords = "social network, Covid-19, large language models (LLMs), social media",

author = "Zhijin Guo and Simpson, \{Edwin D.\} and Roberta Bernardi",

year = "2024",

month = jul,

day = "4",

language = "English",

note = "epiDAMIK 2024: : The 7th International workshop on Epidemiology meets data Miningand Knowledge discovery, epiDAMIK ; Conference date: 26-08-2024 Through 26-08-2024",

}

TY - CONF

T1 - Medfluencer

T2 - epiDAMIK 2024:

AU - Guo, Zhijin

AU - Simpson, Edwin D.

AU - Bernardi, Roberta

PY - 2024/7/4

Y1 - 2024/7/4

N2 - In our study, we first constructed a dataset from the tweets ofthe top 100 medical influencers with the highest Influencer Score[14] during the COVID-19 pandemic. This dataset was then usedto construct a socio-semantic network, mapping both their identities and key topics, which are crucial for understanding theirimpact on public health discourse. To achieve this, we developeda few-shot multi-label classifier to identify influencers and theirnetwork actors’ identities, employed BERTopic for extracting thematic content, and integrated these components into a networkmodel to analyze their impact on health discourse. To ensure thereproducibility of our results, we have made the code available athttps://github.com/ZhijinGuo/Medinfluencer.

AB - In our study, we first constructed a dataset from the tweets ofthe top 100 medical influencers with the highest Influencer Score[14] during the COVID-19 pandemic. This dataset was then usedto construct a socio-semantic network, mapping both their identities and key topics, which are crucial for understanding theirimpact on public health discourse. To achieve this, we developeda few-shot multi-label classifier to identify influencers and theirnetwork actors’ identities, employed BERTopic for extracting thematic content, and integrated these components into a networkmodel to analyze their impact on health discourse. To ensure thereproducibility of our results, we have made the code available athttps://github.com/ZhijinGuo/Medinfluencer.

KW - social network

KW - Covid-19

KW - large language models (LLMs)

KW - social media

M3 - Conference Paper

Y2 - 26 August 2024 through 26 August 2024

ER -

Medfluencer: A Network Representation of Medical Influencers’ Identities and Discourse on Social Media

Abstract

Workshop

Research Groups and Themes

Keywords

Access to Document

Handle.net

Fingerprint

Projects

“Medfluencer: a socio-semantic network analysis method to investigate how medical influencers affect health discourse on social media

Prizes

University of Bristol Research Fellowship

Activities

Invited Talk on “Who says What? Identities and NLI for Studying Health Misinformation on Social Media”

Invited Talk on “Who says What? Identities and NLI for Studying Health Misinformation on Social Media,” WORDS Network Autumn Meeting.

Cite this