Digital Methods Winter School 2017 on “Data Infrastructures: Database Stories, Dumps and Query Driven Narratives”, University of Amsterdam, 9-13th January 2017

I’m involved in co-organising the 2017 edition of the annual Digital Methods Initiative Winter School which is on the theme of “Data Infrastructures: Database Stories, Dumps and Query Driven Narratives”.

I’ll be speaking at the event alongside Professor Geoffrey Bowker (School of Information and Computer Sciences, University of California, Irvine), who is one of the pioneers of the social study of information infrastructures.

If you’re interested in how digital methods can be used to study or tell stories with data infrastructures, then we’d love to see you there. Further details about the theme are available on the Digital Methods Initiative wiki (excerpt copied below).

The Digital Methods Initiative (DMI), Amsterdam, is holding its annual Winter School on Data Infrastructures. The format is that of a (social media and web) data sprint, with hands-on work for telling stories with data, together with a programme of keynote speakers and a Mini-conference, where PhD candidates, motivated scholars and advanced graduate students present short papers on digital methods and new media related topics, and receive feedback from the Amsterdam DMI researchers and international participants. Participants need not give a paper at the Mini-conference to attend the Winter School. For a preview of what the event is like, you can view short video clips from previous editions of the Summer School in 2015 and 2014.

The DMI Winter School is pleased to have Geoffrey Bowker (University of California, Irvine) give the opening keynote. He is author (among other works) of Memory Practices in the Sciences and (with Susan Leigh Star) Sorting Things Out: Classification and its Consequences, both published by MIT Press. Data infrastructures provide the conditions of possibility for social action as well as ways of seeing the world. Among them online data infrastructures these days range widely from social media API query environments as Facebook’s and Twitter’s and secrets repositories and dumps as Wikileaks to interactive databases of missing migrants, uncounted police killings as well as war deaths by social researchers and leading newspapers such as the New York Times and the Guardian. Beneath them are data collection regimes with multifarious goals such as corporate data science, state data transparency and investigative data journalism. These data infrastructures have in common with ‘information infrastructures’ studied by G. Bowker and S. Leigh Star often enormous assemblages of socio-epistemological work invisible to the “the user-at-terminal”. The entire project of scanning the library books and putting into place the query infrastructure, the n-gram viewer, of Google Books (to mention another data infrastructure Bowker also pointed to) has been called ‘infrastructuring,’ which may be mapped out with considerable effort. Indeed, certain of the data collection work — whether vast and automated, laborious and manual and/or stealthy — as well as its ‘databasing’ have been visualised in a form of deconstruction that strives to demonstrate the crucial choices about what to collect and make available to the web browser user. For example, Facebook no longer makes friends data accessible, so as to enhance user privacy but it also forestalls research opportunities such as a like analysis of Donald Trump’s friends. This is one contribution digital methods may make to data infrastructure studies by providing a critical diagnostics of infrastructure by examining the data fields available and outputted by the query machine, and the limitations inhering therein. Researchers may reverse engineer the query design and initial outputs, as was the case with the studies of the ICWatch database (on surveillance workers) and the JD database (concerning Fukushima). In an exploration of the ICWatch database, an activist project that sourced intelligence workers’ profiles from the social networking sites, LinkedIn and Indeed, researchers also provided network-analytical techniques to clean the database, making the open secrets more credible but also created a typical profile of the surveillance worker. In the Fukushima project researchers found with the use of an historical tweet collection-maker that to check and enrich the (limited) Twitter data set about the Fukushima debates would cost over $10,000.

Apart from such critical diagnostics, or the identification of the mechanisms behind the outputs served, digital methods may also repurpose original or typical uses of the databases, and re-narrate the data space and thus the kind of stories they may tell. Stories told from Wikileaks data, for example, often concern how the release of the confidential is endangering or benefits certain states. Indeed one recent narrative (in the New York Times) has it that the leaks benefit the Russian government. Could Wikileaks be put to uses that Julian Assange once called ’scientific journalism’ or tell data stories of other kinds? In one brief study researchers found that Wikileaks data (Afghan warlogs) is rarely used by journalists and bloggers, hardly linking to the original leak as Assange once envisaged. When stories were told, they typically were scandalous, national stories (e.g., supposed military cover-ups). The 2017 Digital Methods Winter School critiques and repurposes data infrastructures and dumps online so as to re-narrate their current dominant uses.


Back to posts