Digital drug consumers - The production of consumption practices in online communities of illicit drug users

这项研究的设计和资助与以下EPSRC研究主题和子主题有关：数字经济;信任，身份和安全;数字商业模式;内容创建和消费。自从20世纪80年代互联网通信技术为发达国家的普通公众所用以来，来自世界这一地区的非法药物使用者通过创建大规模在线通信平台占用了这些资源，这些平台是生产和传播非法药物消费做法的有影响力的数字中心。过去十年的研究已经确定了这些平台的潜力，除其他事项外-增加/减少特定物质的供应和需求，以及增加/减少与特定物质相关的身体、精神和社会危害。与此同时，互联网通信技术已经传播到更广泛的公众，已经与正常的社会活动更加交织在一起，并且在技术意义上取得了显着进步。EPSRC资助的这项研究的目的是通过两种重要的方式来解释这些重大发展：-通过分析来自在线药物消费文化的最新迭代的数据，例如Reddit等主流平台上的大型子论坛和加密在线药物市场（即“暗网市场”或“加密市场”）托管的论坛-通过开发一种新的数字研究方法，该方法采用在计算机科学中开发的分析方法，但现在开始更广泛地应用。这种方法结合了在线数据收集、社交网络分析和自然语言处理，从而能够以前所未有的规模（即在多个平台上进行数百万次观察）对在线非法药物平台的结构和内容进行分析。这导致了对在线毒品消费者文化的高度现代化分析，旨在通过解决以下研究问题，提供关于数字领域如何产生毒品买卖行为的最新情况：-在线市场和社区如何产生/减少对特定药物的需求？网上市场和社区传播了哪些毒品买卖态度和做法？例如，安全使用或实验？社区和市场结构以及平台促进的互动类型如何影响上述内容？这些问题将针对三种类型的平台进行解决-加密市场论坛（即“暗网市场”或“加密市场”），与较新的大规模流行的主流平台（例如Reddit和Facebook）和历史上有影响力的专业讨论论坛相比。作为本研究的一部分，通过应用和开发的定量数字方法，可以对这些平台的大型数据集进行比较分析。这一方法包括自动在线数据收集（爬行和抓取）、描述性统计、社会网络分析/统计和自然语言处理算法（主题建模和词嵌入）的组合。开发这样的方法允许在线社区的社会语义分析，这是能够占的社会结构和主题内容（以及这些的相互作用）在网上互动的规模，这是以前从未见过的在线药物研究领域，同时也有应用程序在这个领域之外，从而有助于在线社会媒体研究一般。

This research was designed and funded in relation to the following EPSRC research themes and sub themes: Digital economy; trust, identity and security; digital business models; content creation and consumption. Ever since internet communication technologies became accessible to the general public of developed countries in the 1980's, illicit drug users from this part of the world have appropriated them by creating massive online communication platforms which function as influential digital hubs for the production and dissemination of illicit drug consumption practices.Research from the previous decade has already established the potential of these platforms to-among other things-increase/decrease supply of and demand for specific substances and increase/decrease physical, mental and social harms associated with specific substances. In the meantime, internet communication technologies have spread to an even larger public, have become even more inter-weaved with normal social activities and have advanced significantly in a technological sense. The aim of this EPSRC-funded research is to account for these significant developments in two important ways:- By analysing data from the most recent iterations of online drug consumer culture such as massive sub-forums on mainstream platforms like reddit and forums hosted by encrypted online drug markets (ie 'darknet markets' or 'cryptomarkets')- By developing a novel digital research methodology that employs analytic methods which were developed within computer science but are now beginning to be applied more widely. This methodology combines online data collection, social network analysis and natural language processing and thus makes it possible to conduct an analysis of the structure and content of online illicit drug platforms on an unprecedented scale (i.e. millions of observations across multiple platforms). This leads to a highly contemporary analysis of online drug consumer culture which aims to provide an up-to-date picture of how drug buying and selling practices are produced in the digital domain by addressing the following research questions:-How is demand for specific drugs produced/reduced in online markets and communities?-Which attitudes and practices of drug buying/selling are being disseminated in online markets and communities? E.g. safe use or experimentation?-How does the community and market structure and type of interaction facilitated by the platform impact the above?These questions will be addressed for three types of platforms - encrypted market forums (i.e. 'darknet markets' or 'cryptomarkets')compared to newer massively popular mainstream platforms (e.g. reddit and Facebook) and historically influential specialised discussion forums. A comparative analysis of large data sets from these platforms is made possible by the quantitative digital methodology which is applied and developed as part of this research. This methodology consists in a combination of automated online data collection (crawling and scraping), descriptive statistics, social network analysis/statistics and natural language processing algorithms (topic modelling and word embeddings). Developing such an approach allows for socio-semantic analysis of online communities which is able to account for both the social structure and thematic content (as well as the interplay of these) in online interactions on a scale which has not been seen before within the field of online drug research, while also having applications outside of this field and thus contributing to online social media research in general.