Dataset: 2,122 Images from a Public Telegram Channel (Jan–Dec 2024) — Online Misogyny (Encrypted / Restricted Access)

This repository contains an encrypted dataset of 2,122 images collected from a public Telegram channel, covering the period January–December 2024. The dataset was produced as part of the research presented at the 49th ANPOCS Annual Meeting, in the paper: “Visual Patterns of Online Misogyny: A Computer Vision Analysis of Telegram Chats”(Portuguese title: Visões Computacionais da Misoginia Online: Uso de IA para Análise de Imagens em Chats do Telegram) The dataset supports research on online misogyny, image-based hate speech, and multimodal digital methods, particularly within social science approaches to large-scale visual content. The accompanying study used computer vision techniques to process and cluster the images in order to identify recurring visual and narrative patterns associated with misogynistic content. The analysis yielded a preliminary visual typology, with clusters that include (among others): memes, screenshots from other platforms, cartoons, attacks on physical appearance, and sexualization of women’s bodies. This dataset is intended to facilitate replication, comparative research, and methodological development in computational social science and digital sociology, especially for researchers investigating multimodal extremist discourse and gendered hate in digital environments. Data protection and access conditions For data protection and image rights reasons, all files in this repository are encrypted.Researchers interested in accessing the decrypted dataset must: Contact the research team via email: [e-mail] Sign a confidentiality agreement / non-disclosure term Receive the password required to decrypt the files No decrypted images are made publicly available in this repository. Authors Raíssa Jeanine Nothaft (UFBA) — ORCID: 0000-0002-3933-0379 Leonardo Fernandes Nascimento (UFBA) — ORCID: 0000-0003-2929-1115 Leonardo Thibau Teixeira de Moraes (UFBA) — ORCID: 0000-0002-3908-6580 Rosana Silva Moore Wedderburn (UFBA) — ORCID: 0000-0001-6839-8762 Ruan Arthur Lima Santos (UFBA) — ORCID: 0009-0003-1358-4690 Contact For data access requests (confidentiality agreement + password), write to: [e-mail] Keywords Telegram; online misogyny; hate speech; computer vision; multimodal analysis; image clustering; social media research; digital sociology. Suggested citation Nothaft, R. J., Nascimento, L. F., Moraes, L. T. T., Wedderburn, R. S. M., & Lima Santos, R. A. (2026). Dataset: 2,122 images from a public Telegram channel (Jan–Dec 2024) — Online Misogyny (Encrypted / Restricted Access). Zenodo.本仓库包含一份加密数据集，内含2122张从公开Telegram频道采集的图像，采集时段为2024年1月至12月。本数据集是发表于第49届ANPOCS年度会议的研究成果的一部分，相关论文为《在线厌女症的视觉模式：Telegram聊天的计算机视觉分析》（葡萄牙语标题：Visões Computacionais da Misoginia Online: Uso de IA para Análise de Imagens em Chats do Telegram）。 本数据集可支持针对在线厌女症、基于图像的仇恨言论以及多模态数字方法的研究，尤其适用于采用社会科学范式开展的大规模视觉内容研究。配套研究采用计算机视觉技术对图像进行处理与聚类，以识别与厌女症内容相关的反复出现的视觉与叙事模式。该分析构建了一套初步的视觉类型学，其聚类类别包括但不限于：表情包、其他平台的截图、卡通漫画、针对外貌的攻击以及对女性身体的性化表述。 本数据集旨在推动计算社会科学与数字社会学领域的可复现研究、比较研究及方法学发展，尤其面向研究数字环境下多模态极端主义话语与性别仇恨的研究者。 数据保护与获取条件 出于数据保护与图像版权考量，本仓库内所有文件均已加密。有意获取解密后数据集的研究者需完成以下要求：通过电子邮箱[e-mail]联系研究团队；签署保密协议/保密条款；获取解密文件所需的密码。 本仓库不公开任何解密后的图像文件。 作者 拉伊萨·吉内·诺夫特（Raíssa Jeanine Nothaft，巴伊亚联邦大学（Universidade Federal da Bahia，简称UFBA））——ORCID：0000-0002-3933-0379 莱昂纳多·费尔南德斯·纳斯西门托（Leonardo Fernandes Nascimento，巴伊亚联邦大学）——ORCID：0000-0003-2929-1115 莱昂纳多·蒂博·特谢拉·德·莫赖斯（Leonardo Thibau Teixeira de Moraes，巴伊亚联邦大学）——ORCID：0000-0002-3908-6580 罗萨娜·席尔瓦·穆尔·韦德伯恩（Rosana Silva Moore Wedderburn，巴伊亚联邦大学）——ORCID：0000-0001-6839-8762 鲁安·阿瑟·利马·桑托斯（Ruan Arthur Lima Santos，巴伊亚联邦大学）——ORCID：0009-0003-1358-4690 联系方式 如需申请数据访问（含签署保密协议及获取密码），请发送邮件至：[e-mail] 关键词 Telegram；在线厌女症；仇恨言论；计算机视觉；多模态分析；图像聚类；社交媒体研究；数字社会学 推荐引用格式 诺夫特, R. J., 纳斯西门托, L. F., 莫赖斯, L. T. T., 韦德伯恩, R. S. M., & 利马·桑托斯, R. A. (2026). 数据集: 来自公开Telegram频道的2122张图像（2024年1月-12月）——在线厌女症（加密/限权访问）. Zenodo.