Socially reconstructing history: The Social History Timestream application.
Tim Fawnsa[*], Sian Bayneb, Jen Rossb, Stuart Nicolc, Ethel Quayled, Hamish Macleodb and Karen Howiee
aCentre for Medical Education, University of Edinburgh, Edinburgh, UK; bMoray House School of Education, University of Edinburgh, Edinburgh, UK; cInformation Services, University of Edinburgh, Edinburgh, UK; dSchool of Health in Social Science, University of Edinburgh, Edinburgh, UK; eSchool of History, Classics and Archaeology, University of Edinburgh, Edinburgh, UK
Abstract
For centuries, print media controlled by powerful gatekeepers have played a dominant part in the recording and construction of history. Digital media open up new opportunities for the social construction of historical narratives that reveal personal and situatedviewpoints. In January 2012, work began at the University of Edinburgh on the design, development and distribution of a web-based Social History Timestream application for social history research projects across a range of disciplines. The application enables researchers to establish dynamically-generated timelines (divided into days, months, years, decades, etc.), to which researchers and members of the public can post photographs, textual descriptions and other media. With the addition of meta-data such as tags and locations, the resulting timelines provide a way to compare thematically-related events across time.
A primary aim of the application is to provide opportunities for researchers to discover serendipitous time-based connections between topics and events that might not previously have been considered. Key to the project’s success will be an engaging interface that allows visitors to see public imagery (e.g. items from the news) alongside personal imagery (e.g. what a given person was doing on that day), organised by themes (e.g. geography, health, politics or media). Among other things, the interface will allow comparison ofmainstream versions of particular themed histories with the personal accounts of those who experienced them, or to visualise the development of ideas, technologies, and social categorisations over time.
At the time of writing, the Timestream application is still in development and is being piloted with three research projects. This paper will focus on one of these – a history of photography practices – to describe emerging theoretical and methodological design considerations, demonstrate the interface and offer insights into the process of using the Timestream application.
Keywords: history; timeline; social research; dynamic; user-generated; social construction
Introduction
Since the invention of the printing press, personal, oral, domestic and local histories have been increasingly dominated in many societies by historical discourses controlled by print media concerns (Lebvre and Martin 1976). Through selection of the content and tone of their mass-produced and widely-distributed texts, groups such as book and newspaper publishers have acted as gatekeepers of the information that people come to think of as the important events of the past. This has been enacted in extreme, systematic ways such as post-war textbook control in Germany and Japan (Hein and Selden 2000), as well as in subtle ways such as disproportionate coverage of blue and white-collar crime in the United States (Graber 1980). The replicability and precision of typography have given printed information the appearance of ‘cold, non-human, facts’ (Ong 2002, 120) and ‘provided a vast new memory for past writings that made a personal memory inadequate.’(McLuhan2005, 189). Not only have print mediaexerted significant influence over the perspectives of the members of societies, they haveresulted in increased homogeneity of collective memory across large populations (McLuhan 2005).
The emergence of digitalmedia has created an opportunity for a broader range of perspectives to be included in new, socially-constructed historical narratives. What is known as ‘participatory culture’ (Jenkins 1992) has come to be closely associated with the capacity of read-write web technologies to enable social media, user-generated content and a range of other participatory practices which were previously difficult or impossible to produce or disseminate at scale: ‘One of the most exciting elements of new media is that they allow us to communicate personally within what used to be prohibitively large groups. This blurs the boundary between mass and interpersonal communication in ways that disrupt both’(Baym 2010, 4).
The implications for power dynamics between ‘producers’ and ‘consumers’ has been best studied in a media context, and particularly in relation to political discourse and activism (Shirky 2009). The argument goes that participatory culture ‘increasingly demands room for ordinary citizens to wield … technologies that were once the privilege of capital intensive industries – to express themselves and distribute those creations’ (van Dijck 2009, 42). As van Dijck goes on to argue, this stark polarisation between citizen and industry is overly simplistic, and the utopic vision of a democratising, empowering internet is matched by a dystopic one that sees surveillance, control, and economic and social divisions, as equally likely effects of technological change (Hand 2009). Alongside this are concerns that the scale of information produced through ‘citizen journalism’ creates challenges around information literacy and source credibility (Carlson and Franklin 2011). In any case, the shift of communicative resources from the few to the many is likely to have profound implications for social history.
While there is precedent for user-generated accounts of social history – most notably in the Mass Observation social research organisation of the 1940s and 50s in Britain (Mass Observation 2012) – the web redefines what is meant by ‘mass’, and radically decentralises the gaze of the observer. Social scientists, historians and other researchers are still coming to terms with the implications of the explosion of personal accounts now openly accessible, and able to be added to, by those with even quite modest levels of technical skill and access (Wilson, Gosling and Graham 2012). Recent years have seen a movement towards the privileging of so-called ‘big data’(Boyd and Crawford 2012), and the large-scale quantitative analysis of digital content and interactions. At the same time, there is scope within the social media ecology for a more user-generated, emergent and small-scale approach to social science and social history research. This paper describesthe development of a pilot version of the Social History Timestream Application (‘Timestream’) as an example of how researchers and developers can work together to produce environments conducive to such emergent, user-generated methods for social research. Designed as a free-to-use research tool, Timestream displays dynamically-generated timelines to which research teams and members of the public can post photographs, textual descriptions and other media. Rather than report on an evaluation of the success or otherwise of project outcomes, the purpose of this paper is to use the experience of designing, developing, testing and evaluating this tool to reveal insights about the nature of such tools and their implication for social history construction and the methodological considerations of related research.
The Social History Timestream application
Initially funded by the University of Edinburgh Challenge Investment Fund, the pilot phase of this project (which began in January 2012) is designed to explore the potential of the Timestream application. So far, the project has involved a four-month design phase, followed by a period of development and testing. This has involved building the application, populating the database with data for three research projects (see Use cases below), and building a visualisation interface to display these data. Currently, having consulted with members of the research project teams, some further development work is being carried out in response to their feedback.
Among other things, Timestream allows visitors to seetimelines of mainstream imagery (e.g. items from the news) alongside personal imagery (e.g. what a given person was doing on that day), and can be organised by themes (e.g. location, health information, politics or media). In this way, established versions of particular themed histories can be compared with the personal accounts of those who experienced them and new social constructions of the development of ideas, technologies, and social categorisations can be generated. Figure 1 shows an example of personal accounts of photography practices placed alongside mainstream and commercial events.
Figure 1. about here
Once development is sufficiently advanced, Timestream will enable crowd-sourced data to be collected via social media with the aim of drawing from a more extended and diverse selection of historical events than is possible when data are collected from a smaller number of sources. This idea might be thought of as ‘Historical lifestreaming’ – capturing the evolution of mundane or otherwise, personal and public activities over long periods. Though much of the data collected from social media may pertain to current events, we hope that the value of seeing themes develop over longer periods will encourage people to submit accounts of events that occurred prior to the Web2.0 era. With sufficient data, the combination of many different themes will become possible and it is through this process that we hope to generate unexpected discoveries.
There are a number of projects which relate to one or more aspects of the Timestream application. As an example of social history research, the Mass Observation projects involve the compilation of personal accounts to generate an ‘anthropology of ourselves’ (Mass Observation 2012). Rather than being a social history research project, Timestream is a tool for use in research. Indeed, data from the Mass Observation projects might be usefully entered into the Timestream database for visualisation. As an application, Timestream is more directly comparable to other timeline technologies such as Timeglider, Timekiwi, Memolaneor Dipity see their websites, respectively). These applications, however, construct timelines around single themes rather than facilitate comparison across multiple themes. They are typically designed for display rather than analysis and do not have customisable filters for isolating data according to user-manipulated parameters.
To our knowledge, there is no other tool in existence designed for the sort of comparative, time-based visualisation that Timestream enables. The visualisation interface is being developed using the Simile Widgets Timeline Javascript library (Massachusetts Institute of Technology 2009), an opensource platform containing a large number of required functions which can be adapted to Timestream’s needs. This library was created from the Semantic Interoperability of Metadata and Information in unLike Environments (SIMILE) project at MIT which ‘focused on developing robust, open source tools that empower users to access, manage, visualize and reuse digital assets’ (Massachusetts Institute of Technology 2008). The use of open source software aligns well with the inclusive philosophy of the Timestream project and we plan to make our software opensource if further funding is secured after the pilot phase.
TheSimile Widgets Timeline library is integrated with PHP and MySQL to allow users to interact with a database of events that is designed to be compatible with event-based data used in other online history projects. This should allow Timestream to import data from accessible sources and to export data (where permission is given by the appropriate party) to other platforms. Each research project within Timestream may be open to the public or restricted to members of the research team. Within this, events belonging to each project may be further restricted. This enables researchers to protect information that is confidential while allowing non-confidential data to be used in other projects.
Use cases
Engaging with real research projects during the pilot phase helped us to build our design specifications around what is useful in practice. Due to time constraints, we started by building data entry forms to allow researchers to contribute events while we were developing the visual interface. This overlap allowed us to tweak our design as and when unforeseen needs or challenges arose. Three research projects based at the University of Edinburgh were used as test cases. The benefit to the projects was the possibility that visualising many discrete events over time could inform further analysis by revealing insights and additional questions. Although data for all of these projects have initially been entered by the research team, the eventual aim is for members of the public to contribute their own experiences via social media interfaces. The effect that this opening up of the data might have is considered under the Methodological Implicationssection later in the article.
The projects included the Nighttime Dementia Care project, the Theology and Therapy project (which explored connections between psychotherapy, Christianity and the language of spirituality in post-war Scotland and England) and the History of Photography Practices project. This last project, which explores how people have captured, organised, reviewed and shared photographs over time, is the focus of the remainder of this article. By adding events that relate to technological change or adoption, personal accounts (taken from interviews and previously-published case studies) are contrasted with mainstream documentation of technological advances. Led by a member of the Timestream team, this work has provided particularly illuminating experiences of engaging with both the development of the application and with the process of using it in research. As such, it has been invaluable in increasing our understanding of the likely requirements and challenges that other researchers are likely to face when using Timestream. Taking the History of Photography Practices project as a case study, an explanation is given below of the sort of data Timestream holds and how it is visualised, followed by a discussion of practical, conceptual and methodological issues that emerged during design, development and testing of the application.
Entering and viewing data
The Timestream database is comprised of event objects, each of which contains attributes controlled by the event ‘owner’ (the researcher or member of the public who enters it into the database) such as the date of the event, the interval level to which the event can be specified (i.e. day, week, month, year, decade or century), associated images and descriptions.Events are also assigned tags which are precise, abbreviated analytic codings such as ‘privacy’ or ‘automation’. Sets of tags that comprise a theme are used to create ‘categories’ which act as filters for determining which events to display. An example of a category from the History of Photography Practices project is ‘practices’ which consists of the tags ‘capturing’, ‘reviewing’, ‘organising’ and ‘sharing’. Multiple categories can be compared in parallel so that, for example, a timeline of news broadcasts on technological developments could be viewed beside personal accounts reminiscence. Figure 1 (above) demonstrates the categories ‘personal’, ‘official’ and ‘commercial’.
Data are entered in the context of a particular research project such as those outlined under Use Cases above. They may, however, be viewed in various different contexts alongside data from other projects. This enables cross-fertilisation of ideas where events pertaining to one theme can be associated or contrasted with those from other, potentially diverse projects. For example, Muybridge’s (1957) ‘Horse in Motion’, which shows sequential frames of the position of a horse’s legs while galloping, might reveal an important cross-over between advances in photography and video. For this reason, the Timestream viewing interface is configured according to ‘views’rather than projects. A view consists of categories, each of which contains data from a single project but which may be collated from different projects. Views also have parameters such as default display date, date interval level (e.g. day, month, year, etc.), which tags are highlightedwithin the events associated with them, and magnification level (the size at which events will be displayed and, therefore, how many events can be displayed). Each user can save multiple views, meaning that it is possible to return to precisely-specified visualisations and share them with other users. See Figure 2 for an example display of a view and its configuration.
Figure 2 about here.
At the beginning of the pilot, it was difficult to imagine the details of the Timestream visualisation interface. We did not know, for example, whether time should scroll horizontally or vertically, or how to display the essence of an event within a small space. It was not until we had built a simple prototype and begun to enter data that we were able to make such decisions. Likewise, researchers needed to see what visualisations might look like before they felt able to make decisions about what data should be entered and how. Hence, design and data entry became iterative, interdependent processes that informed each other’s development.
To increase the power and flexibility of visualisations, one researcher suggested that alongside filtering by categories (i.e. displaying only those events that have been assigned specified tags), tags could be highlighted within their associated events (see Figure 2). This meant that we could show where tags were present across categories, potentially revealing unexpected relationships. In hindsight, this seems like an obvious requirement but it was not one that we had discussed prior to engaging with a real research project.