Tag Archives: research

The limits of social media big data

handbook social media researchA new book chapter by Rob Kitchin has been published in The Sage Handbook of Social Media Research Methods edited by Luke Sloan and Anabel Quan-Haase. The chapter is titled ‘Big data – hype or revolution’ and provides a general introduction to big data, new epistemologies and data analytics, with the latter part focusing on social media data.  The text below is a sample taken from a section titled ‘The limits of social media big data’.

The discussion so far has argued that there is something qualitatively different about big data from small data and that it opens up new epistemological possibilities, some of which have more value than others. In general terms, it has been intimated that big data does represent a revolution in measurement that will inevitably lead to a revolution in how academic research is conducted; that big data studies will replace small data ones. However, this is unlikely to be the case for a number of reasons.

Whilst small data may be limited in volume and velocity, they have a long history of development across science, state agencies, non-governmental organizations and business, with established methodologies and modes of analysis, and a record of producing meaningful answers. Small data studies can be much more finely tailored to answer specific research questions and to explore in detail and in-depth the varied, contextual, rational and irrational ways in which people interact and make sense of the world, and how processes work. Small data can focus on specific cases and tell individual, nuanced and contextual stories.

Big data is often being repurposed to try and answer questions for which it was never designed. For example, geotagged Twitter data have not been produced to provide answers with respect to the geographical concentration of language groups in a city and the processes driving such spatial autocorrelation. We should perhaps not be surprised then that it only provides a surface snapshot, albeit an interesting snapshot, rather than deep penetrating insights into the geographies of race, language, agglomeration and segregation in particular locales. Moreover, big data might seek to be exhaustive, but as with all data they are both a representation and a sample. What data are captured is shaped by: the field of view/sampling frame (where data capture devices are deployed and what their settings/parameters are; who uses a space or media, e.g., who belongs to Facebook); the technology and platform used (different surveys, sensors, lens, textual prompts, layout, etc. all produce variances and biases in what data are generated); the context in which data are generated (unfolding events mean data are always situated with respect to circumstance); the data ontology employed (how the data are calibrated and classified); and the regulatory environment with respect to privacy, data protection and security (Kitchin, 2013, 2014a). Further, big data generally capture what is easy to ensnare – data that are openly expressed (what is typed, swiped, scanned, sensed, etc.; people’s actions and behaviours; the movement of things) – as well as data that are the ‘exhaust’, a by-product, of the primary task/output.

Small data studies then mine gold from working a narrow seam, whereas big data studies seek to extract nuggets through open-pit mining, scooping up and sieving huge tracts of land. These two approaches of narrow versus open mining have consequences with respect to data quality, fidelity and lineage. Given the limited sample sizes of small data, data quality – how clean (error and gap free), objective (bias free) and consistent (few discrepancies) the data are; veracity – the authenticity of the data and the extent to which they accurately (precision) and faithfully (fidelity, reliability) represent what they are meant to; and lineage – documentation that establishes provenance and fit for use; are of paramount importance (Lauriault, 2012). In contrast, it has been argued by some that big data studies do not need the same standards of data quality, veracity and lineage because the exhaustive nature of the dataset removes sampling biases and more than compensates for any errors or gaps or inconsistencies in the data or weakness in fidelity (Mayer-Schonberger and Cukier, 2013). The argument for such a view is that ‘with less error from sampling we can accept more measurement error’ (p.13) and ‘tolerate inexactitude’ (p. 16).

Nonetheless, the warning ‘garbage in, garbage out’ still holds. The data can be biased due to the demographic being sampled (e.g., not everybody uses Twitter) or the data might be gamed or faked through false accounts or hacking (e.g., there are hundreds of thousands of fake Twitter accounts seeking to influence trending and direct clickstream trails) (Bollier, 2010; Crampton et al., 2012). Moreover, the technology being used and their working parameters can affect the nature of the data. For example, which posts on social media are most read or shared are strongly affected by ranking algorithms not simply interest (Baym, 2013). Similarly, APIs structure what data are extracted, for example, in Twitter only capturing specific hashtags associated with an event rather than all relevant tweets (Bruns, 2013), with González-Bailón et al. (2012) finding that different methods of accessing Twitter data – search APIs versus streaming APIs – produced quite different sets of results. As a consequence, there is no guarantee that two teams of researchers attempting to gather the same data at the same time will end up with identical datasets (Bruns, 2013). Further, the choice of metadata and variables that are being generated and which ones are being ignored paint a particular picture (Graham, 2012). With respect to fidelity there are question marks as to the extent to which social media posts really represent peoples’ views and the faith that should be placed on them. Manovich (2011: 6) warns that ‘[p]eoples’ posts, tweets, uploaded photographs, comments, and other types of online participation are not transparent windows into their selves; instead, they are often carefully curated and systematically managed’.

There are also issues of access to both small and big data. Small data produced by academia, public institutions, non-governmental organizations and private entities can be restricted in access, limited in use to defined personnel, or available for a fee or under license. Increasingly, however, public institution and academic data are becoming more open. Big data are, with a few exceptions such as satellite imagery and national security and policing, mainly produced by the private sector. Access is usually restricted behind pay walls and proprietary licensing, limited to ensure competitive advantage and to leverage income through their sale or licensing (CIPPIC, 2006). Indeed, it is somewhat of a paradox that only a handful of entities are drowning in the data deluge (boyd and Crawford, 2012) and companies such as mobile phone operators, app developers, social media providers, financial institutions, retail chains, and surveillance and security firms are under no obligations to share freely the data they collect through their operations. In some cases, a limited amount of the data might be made available to researchers or the public through Application Programming Interfaces (APIs). For example, Twitter allows a few companies to access its firehose (stream of data) for a fee for commercial purposes (and have the latitude to dictate terms with respect to what can be done with such data), but with a handful of exceptions researchers are restricted to a ‘gardenhose’ (c. 10 percent of public tweets), a ‘spritzer’ (c. one percent of public tweets), or to different subsets of content (‘white-listed’ accounts), with private and protected tweets excluded in all cases (boyd and Crawford, 2012). The worry is that the insights that privately owned and commercially sold big data can provide will be limited to a privileged set of academic researchers whose findings cannot be replicated or validated (Lazer et al., 2009).

Given the relative strengths and limitations of big and small data it is fair to say that small data studies will continue to be an important element of the research landscape, despite the benefits that might accrue from using big data such as social media data. However, it should be noted that small data studies will increasingly come under pressure to utilize the new archiving technologies, being scaled-up within digital data infrastructures in order that they are preserved for future generations, become accessible to re-use and combination with other small and big data, and more value and insight can be extracted from them through the application of big data analytics.

Rob Kitchin

New paper in Information, Communication and Society: Thinking critically about and researching algorithms

A new paper by Rob Kitchin – Thinking critically about and researching algorithms – has just been published (online first) in Information, Communication and Society.


More and more aspects of our everyday lives are being mediated, augmented, produced and regulated by software-enabled technologies. Software is fundamentally composed of algorithms: sets of defined steps structured to process instructions/data to produce an output. This paper synthesises and extends emerging critical thinking about algorithms and considers how best to research them in practice. Four main arguments are developed. First, there is a pressing need to focus critical and empirical attention on algorithms and the work that they do given their increasing importance in shaping social and economic life. Second, algorithms can be conceived in a number of ways – technically, computationally, mathematically, politically, culturally, economically, contextually, materially, philosophically, ethically – but are best understood as being contingent, ontogenetic and performative in nature, and embedded in wider socio-technical assemblages. Third, there are three main challenges that hinder research about algorithms (gaining access to their formulation; they are heterogeneous and embedded in wider systems; their work unfolds contextually and contingently), which require practical and epistemological attention. Fourth, the constitution and work of algorithms can be empirically studied in a number of ways, each of which has strengths and weaknesses that need to be systematically evaluated. Six methodological approaches designed to produce insights into the nature and work of algorithms are critically appraised. It is contended that these methods are best used in combination in order to help overcome epistemological and practical challenges.

The final version is available here and a pre-print version of the paper can be downloaded here.

New paper: Thinking critically about and researching algorithms

A new paper by Rob Kitchin has been posted as open access on SSRNThinking critically about and researching algorithms is The Programmable City Working Paper 5.

The era of ubiquitous computing and big data is now firmly established, with more and more aspects of our everyday lives being mediated, augmented, produced and regulated by digital devices and networked systems powered by software.  Software is fundamentally composed of algorithms — sets of defined steps structured to process instructions/data to produce an output.  And yet, to date, there has been little critical reflection on algorithms, nor empirical research into their nature and work.  This paper synthesises and extends initial critical thinking about algorithms and considers how best to research them in practice.  It makes a case for thinking about algorithms in ways that extend far beyond a technical understanding and approach.  It then details four key challenges in conducting research on the specificities of algorithms — they are often: ‘black boxed’; heterogeneous, contingent on hundreds of other algorithms, and are embedded in complex socio-technical assemblages; ontogenetic and performative; and ‘out of control’ in their work.  Finally, it considers six approaches to empirically research algorithms: examining source code (both deconstructing code and producing genealogies of production); reflexively producing code; reverse engineering; interviewing designers and conducting ethnographies of coding teams; unpacking the wider socio-technical assemblages framing algorithms; and examining how algorithms do work in the world.

Key words: algorithm, code, epistemology, research