Leeds Critical Data Studies Group

Schedule

Seminars usually take place on the last Wednesday of every month and generally take place at the Leeds Institute for Data Analytics (LIDA) (located on the 11th floor of the Worsley Building) in Boardroom 11.87, but there are one or two changes to this so please see the schedule below.

Wednesday 28 September 2016, 3.30-5.00pm
The social life of algorithms
Dr Carlo Perrotta, University Academic Fellow in Digital Learning, School of Education, University of Leeds
Venue: LIDA, Boardroom 11.87

Carlo Perrotta will be talking about the ‘social life’ of algorithms, discussing the case of learning analytics in education based on a recently published paper that he has written with Ben Williamson [1]. Carlo will argue that analytic techniques are seldom (if ever) ‘pure’ instruments in the hands of neutral actors, and that the operations procedurally performed on the data conflate the representation and the construction of the phenomena under study. The related paper draws on material semiotics and performativity to examine one specific algorithmic method (cluster analysis) as a ‘performative device’ that, to a significant extent, creates the educational entities it claims to objectively represent.

[1] http://www.tandfonline.com/doi/full/10.1080/17439884.2016.1182927

Wednesday 26 October 2016, 3.30-5.00pm
What’s next for Critical Data Studies?
Dr Heather Ford, University Academic Fellow in Digital Methods, School of Media and Communication and Leeds Institute for Data Analytics, University of Leeds
Venue:LIDA, Boardroom 11.87

In this session, Heather Ford for will provide an overview of some of the main concerns and key thinkers in the areas of critical data studies, algorithmic studies, software studies and related areas. What is there to be critical about when it comes to digital data and the socio-technical arrangements that have been reconfigured by the surge of interest and practice in data collection and analysis? What is the value of such critique? And where should we at the University of Leeds be heading when it comes to making research interventions in this area? The session will begin with a presentation but will be followed by an open discussion about the university’s role in the field of critical data studies more specifically.

Wednesday 23 November 2016, 4.00-5.00pm
Investigating Twitter in Australia: Challenges of Large Scale Social Media Analytics
Dr Brenda Moon, Queensland University of Technology
Venue: LIDA, Boardroom 11.09 (note different room)

At the Queensland University of Technology, the Digital Media Research Centre (DMRC) [1] conducts research that helps society understand and adapt to the social, cultural and economic transformations associated with digital media technologies. Dr Brenda Moon [2] contributes to an Australian Research Council (ARC) Future Fellowship project, led by Professor Axel Bruns [3], which draws on large datasets and innovative methods to develop a new model of the Australian online public sphere. In this presentation Brenda will introduce the DMRC and describe how the Australian online public sphere project is using the comprehensive Australian TrISMA Twitter and Facebook datasets to develop methods which address the challenges and opportunities of working with large scale social media analytics. She will present the example of how the TrISMA dataset has been used to address one limitation of hashtags studies; that they only consider the parts of the conversation which contain the hashtag. The TrISMA dataset allows us to recursively extend this to include the conversation that is in reply to, or is replied to by, the conversation identified using hashtags.

[1] QUT DMRC: http://www.qut.edu.au/research/dmrc
[2] Brenda Moon: http://staff.qut.edu.au/staff/moonb/
[3] Axel Bruns: http://staff.qut.edu.au/staff/bruns/

Wednesday 30 November 2016, 3.30-5.00pm
Taking Shortcuts: Correlation not Causation, and the Moral Problems it Brings
Dr Kevin Macnish, IDEA Centre, University of Leeds
Venue: LIDA, Boardroom 11.87

Large scale data analytics have raised a number of ethical concerns.  Many of these were introduced in a seminal paper by boyd and Crawford in 2012 and have been developed since by others (boyd and Crawford 2012; Martin 2015; Lagoze 2014).  One such concern which is frequently recognized but under-analysed is the focus on correlation of data rather than on the causative relationship between data and results.  Advocates of this approach dismiss the need for an understanding of causation, holding instead that the correlation of data is sufficient to meet our needs.  In crude terms, this position holds that we no longer need to know why X+Y=Z.  Merely acknowledging that the pattern exists is enough. In this paper I explore the ethical implications and challenges surrounding a focus on correlation over causation.  In particular, I focus on questions of legitimacy of data collection, the embedding of persistent bias, and the implications of future predictions.  My conclusion is that by failing to consider causation, the short-term benefits of speed and cost may be countered by ethically problematic scenarios in both the short and long term.

Wednesday 22 February 2017, 3.30-5.00pm
Big Data, Big Danger!
Dr Richard P. Mann, University Academic Fellow in Data Analytics, School of Mathematics, University of Leeds
Venue: LIDA, Boardroom 11.87

New data analytic methods, combined with large data sets and new types of data, offer the possibility to quantitatively study human behaviour with unprecedented resolution. At the same time, these new methods and data present clear dangers to the working scientist. This talk will focus on how we can be misled in an age of Big Data and Big Analytics, and what we can do to ensure our research is robust and repeatable.  

Wednesday 29 March 2017, 3.30-5.00pm
Using police data to research Human Trafficking
Dr Carly Lightowlers (University of Leeds, School of Law) and Dr Rosemary Broad (University of Manchester)
Venue: LIDA, Boardroom 11.87

To date, there is considerable less research and available data on the offenders of Human Trafficking than the victims. Police and other agencies routinely collect intelligence and crime, arrests and charges and these contain details of the victims, offenders and key locations in which these interact and operate. However, such administrative data are held in a number of different IT systems; meaning the data are fragmented across a number of sources. Moreover, they are captured differently across different agencies and indeed police services meaning there are significant issues of quality concerning these data. Some key challenges include: extracting meaningful data for (academic) analyses and encouraging the routine and systematic collection of data across police forces to enable meaningful comparison. Alongside colleagues from the University of Manchester, work is underway to develop comprehensive data collection to allow for further analysis, including the geographic analysis of the suspected locations of trafficking activity by offence type. Such data may also offer insights the profiles of offenders (as well as victims) using data reduction techniques and classification tools based on machine learning approaches could be developed. The success of such further work, however, hinges on their being enough data of sufficient quality and ethical practice in dealing with these sensitive data. This seminar will explore the challenges, successes and insights gleaned to data from working alongside Greater Manchester and West Yorkshire Police to explore the potential of data on Human Trafficking.

Wednesday 26 April 2017, 3.30-4.30pm
Changing the world with data visualisation … for the worse? Assessing data visualisations in the abortion debate
Dr Rosemary Lucy Hill, School of Sociology and Social Policy
Venue: LIDA, Boardroom 11.87

Data visualisations are argued to have the power to change the world, as they provide accessibility to data, thereby making decisions more rational (Kosara et al., 2009; Few, 2008). This assertion relies upon the common-sense perception of visualisations as windows onto objective data, implicit in which is the ideal that data visualisations can counter non-factual media messages or providing an affective experience. But notions of what counts as ‘good’ are subjective, and just as all of these elements of a visualisation can be harnessed for good, so can they be harnessed to do very bad things. This paper will consider the discourse of data visualisations’ power to ‘do good’ (Periscopic, 2014) within the context of the highly contested abortion debate. It will report on the early stages of the Persuasive Data project which investigates abortion-related visualisations’ persuasive capabilities against the backdrop of political, religious and moral ideologies. It will ask, how are data visualisations being used to persuade people within the abortion debate? And how should we judge which are the ‘bad’ kind of visualisations when so much is at stake?

Thursday the 11 May, 4-5pm
Privacy in Big Data Analytics? Not under Opacity
Dr Vincent C. Müller, University of Leeds (Philosophy) & Anatolia College/ACT
Venue: LIDA, Boardroom 11.09

There are detailed rules and laws for data protection, especially in medical contexts, but even with the best intentions it is impossible to follow these rules in the practice of big data analytics. In big data, the analysts and the subjects of personal data cannot know what information the data might yield, especially if combined with other data: Big data is opaque. The result is that extant rules and laws cannot protect private information.

© Copyright Leeds 2017