Thread by @ponguru

🎉🎉 Our paper “A Virus Has No Religion”: Analyzing Islamophobia on Twitter During the COVID-19 Outbreak, accepted at @ACMHT 🎉🎉 Work with amazing students @mohit__30 @ManvithReddy18 @shradhasgl @reallysaurabh #IIITStudentsRock
Pre-print: arxiv.org/pdf/2107.05104…
Findings 🧵

We study the rise of Islamophobia on Twitter during the pandemic. We make public the CoronaIslam dataset precog.iiitd.edu.in/resources.html with 410,990 tweets from Feb-May’20. 2000 tweets annotated as hateful and non-hateful for a binary Islamophobia-classifier

We find a link in online & offline behaviour with spikes in online tweets occurring around major real-world events. Most hateful tweets were made when news reports linked Tablighi Jamaat congregation to the spread of #COVID19. Graph with temporal spikes 📉 and table of events 👇

We study the change in context of words ‘Muslim’ and ‘Virus’ through the 4 months. We find terms like ‘Virus’ and ‘Covid19’ are used in a semantically similar context as ‘Muslim’ from March onwards! The term ‘Virus’ is also increasingly associated with ‘Muslim’.

Topic modelling yields both hateful & non-hateful topics. Hateful topics revolve around Tablighi Jamaat & blame the Muslim community for spread of virus and attack of health workers. Non hateful topics included the topic of Ramadan when Muslims donated plasma & helped others!🤝

We study user characteristics of users posting hateful & non-hateful tweets. We find an association b/w religion & how one perceives Muslims. Word clouds of bios of users w/ least to most no. of hateful tweets (L-R). Network graph shows separate communities based on hate-posting.

We study popular URLs referenced in CoronaIslam dataset.

We study content from @OpIndia_com @BBC @YouTube. OpIndia articles cover Tablighi Jamaat, BBC articles cover Uighur Muslims & COVID precautions.

Overall, we find a rise in hate sentiments towards Muslims during the pandemic. Shoutout to other works exploring racism associated with COVID.
Sinophobia studied by @zsavvas90 @jhblackb @gianluca_string @ciciling07 @realyangzhang @CISPA @iDRAMALab 🙏🙏

Asian hate studied by @cjziems, @soni_sandeep @srijaankedia @GeorgiaTech 🙏🙏
Our dataset is also obtained by filtering a larger COVID-19 dataset by @thatmchen @KristinaLerman @emilio__ferrara

Tagging organizations that fight hate-speech: @adl @APC_News @hatespeechbeda @officialuarh @nohate_speech @AkyolinEnglish

Paper again 📜 : arxiv.org/pdf/2107.05104…