Conceptual Background

Extract from the 2007 ARC Linkage Application

Context

Maps and timelines are two of the most effective ways of communicating information about the past. They are widely used in printed atlases of history and archaeology, in CD-ROMs and web sites associated with major history textbooks, and in many museums, visitor centres, television programs and public interest and educational web sites. Maps conventionally show where things are located; timelines show when things happened; the two often occur together, but rarely operate in concert.

One might assume that these forms of visualisation are well understood, adequately conceptualised, and easily generated. While this is generally true for maps – which have a long history of cartographic research, a well-established scope and vocabulary, broad conventions and adequate software tools – there is a striking void in the conceptualisation and practical development of timelines. This project aims to fill this void.

Figure 1: A conventional timeline - scrollable, thematic but not zoomable and without event relationships

The majority of current historical timeline applications are either chronological lists of events or simple two-dimensional graphs on which points, lines or icons representing events are distributed along a time axis in one direction, and arranged thematically or by geographic region in the other (Figure 1). In some cases the graph can be panned and zoomed, but the content remains unchanged (zooming-in on the graph makes it more readable, but does not render additional detail). Most timelines are hand-built representations of a fixed set of data, emphasising situation-specific design over reusability. In rare cases the graph is built on-the-fly from a data file or database query, allowing reuse of the timeline design for different situations (Figure 2).

Figure 2: The SemTime project (Jensen 2006), one of the few projects attempting to tackle the semantics of event relationships and new modes of timeline visualisation.
Aims

We aim to: introduce new methods of understanding and representing history; improve the delivery of historical information in museums, on the web and in the classroom; and enable the production of engaging multimedia presentations of historical narrative without the costs (both technical and financial) currently associated with such resources. In doing so we aim to improve the quality of visitor/student experience, and the understanding of history that they acquire.

The use and value of timelines as a means of presenting historical information has been seriously restricted by the failure to connect maps and timelines within a single application; by the limitation of timeline visualisation to the simple 2D charts described above; and by the lack of appropriate tools to build databases of historical events and to generate timelines automatically from those databases.

The specific aims of this project are therefore to:

  1. define, for the first time, the field of historical timeline visualisation and create a vocabulary to describe the functions required within such visualisations. In a sense we will be doing for timelines what 19th century cartographers did for the 2D map: development of a descriptive language and definition of the limits of the concept and nature of the practice;
  2. create a new data model which situates historical events in both time and space, defines the relationships between events and accommodates the uncertainty and diffuse boundaries (both temporal and spatial) which are characteristic of historical information. No existing model addresses more than a subset of these important issues which are essential for the effective creation of map and timeline visualisations;
  3. create a novel timeline visualisation tool which renders the rich context and interrelation of historical events and operates in tight collaboration with time-based maps. The tool will render data from a server database, allowing a single configurable tool to provide timeline visualisations for many different datasets;
  4. develop methods of building content to drive the visualisation tool. These will include 1. a collaborative database, which allows anyone – from content specialists to students and the general public – to enter data, engage in discussions and generate and publish their own timeline and linked map visualisations; and 2. data mining of textual content to identify people, places and events and assign spatial, temporal and thematic metadata to content items;
  5. develop and evaluate pedagogical models for the use of linked map and timeline visualisations in communicating history through museum exhibitions (test case: Australian National Maritime Museum), historical information web sites (test case: MacquarieNet) and in face-to-face teaching (test case: undergraduate history courses at U. of Sydney and U. of California Merced).
Background

There is a huge literature – and practice – on the use of spatial information systems (notably Geographic Information Systems) for contemporary applications dealing with a single (generally contemporary) moment in time. Location-based services and the use of spatial data (eg. Google Maps, satellite navigation, fleet management and emergency response systems, environmental analysis) are a pervasive part of modern society and one of the major growth areas of the Information Technology industry in the 21st century.

Time – in so far as it is considered in spatial systems – is generally handled through a series of chronologically ordered ‘snapshots’. Snapshot data record the state of affairs at particular moments in time rather than tracking the history of individual entities (eg. people, administrative units, trade routes, ideas), and are thus limited in their ability to handle the complexity of historical events and their relationships. Langren’s (1992) seminal book on temporal GIS reveals the complexity of dealing with spatial entities through time and is widely quoted, but only a handful of researchers – primarily originating from Pennsylvania State University – have pursued the issues, and no mainstream or commercial GIS has implemented more than a small subset of spatio-temporal capabilities (generally related to point entities and/or networks, used for object tracking, network traffic control and transport management).

Spatial information systems have been readily adopted in spatially-oriented Humanities disciplines such as Archaeology, to record and manage data, to visualise spatial distributions, and as an analytical platform to analyse spatial relationships between human activities, environment and landscape (see for example Johnson & North 1995, Conolly 2006). Uptake in History has been rather slower, but recent publications such as Knowles (2002, 2005, 2008) show how the adoption of GIS can offer new insight into historical questions eg. Cunfer’s (2002) reassessment of the origin and development of the American Dustbowl.

However, many historical GIS applications focus on the geography of a problem at a given time (for example, a snapshot of a city’s transport network prior to the decline of the tram system, or successive snapshots reflecting different stages in the development of a military campaign) or on the changing attributes of a location (for example, changing population in a settlement or census unit). The changing spatial location, extent and attributes of historical phenomena which persist through time and change continuously (but are observed only at intervals) are much harder to model, analyse or visualise.

A number of recent projects have attempted to model historical events. The best known is the Historical Event Markup Language (Robinson 2004), which supports date stamping and point locations, and generates either a simple timeline or map visualisation (but not together). The Time Period Directory (Petras et al. 2006) originating in the library world, develops a ‘gazeteer’ approach to place and time using textual descriptions. Our own work on TimeMap (Johnson 1999, 2002) goes further in integrating timestamp data with geographic objects (point, line, polygon) and generating time-enabled maps (but not timelines). The best work on simultaneous display of events on maps and timelines probably comes from crime mapping (Chainey & Ratcliffe 2005).

Other recent event modelling projects focus on time to the exclusion of location: the Temporal Modelling Project (Drucker & Novwiskie 2005) explores the nature of events; SemTime (Jensen 2006) focuses on relationships and novel ways of visualizing timelines (Figure 1, right); TimeML (Pustejovsky et al. 2003) has a rich language for expressing dates and intervals. Of all these projects, only the Time Period Directory provides any methods for recording the uncertain, diffuse and multi-vocal nature of historical information, while only SemTime addresses the issues of relationships between events.

Our project will build on these initiatives, integrating their separate directions into a single methodology and a practical toolkit. Our work will focus particularly on the definition of relationships between events (for example: the nesting of detail events within parent events; influence and causality; thematic relationships; and relationships within a narrative eg. a journey, a person’s life, a dynasty or an historical interpretation of a set of events). We regard relationships as the key to historical narrative and context, and an area which has received inadequate attention. Our preliminary work on building an event relationship model and a pilot event database browser has convinced us of the great potential of this approach.

Prior Work

The project is built on a very significant base of existing work by, and collaboration between, the project partners. It draws on the substantial experience of the Archaeological Computing Laboratory (ACL) directed by Johnson, and the Centre for Research on Computer Supported Learning and Cognition (CoCo) co-directed by Reimann, in delivering real-world interactive and collaborative solutions:

  • TimeMap (http://sydney.edu.au/arts/timemap/) developed by Johnson and the ACL since 1997, was the first software to provide generic interactive time-filtered web maps and map animation of historical features. It will provide the core mapping components for this project;
  • Heurist (http://heuristnetwork.org, developed by Johnson and the ACL since 2005, provides a flexible collaborative database environment which will form the core data infrastructure for this project (with modifications);
  • Heurist has already been used by Johnson and Mostern to build a proof-of-concept implementation of an historical event model with spatial and temporal extents and inter-event relationships, funded by a small grant from the Hewlett Foundation;
  • Johnson and Mostern have collaborated since 1998 as key members of the Electronic Cultural Atlas Initiative (http://ecai.org/) which has been pivotal in the development of several of the historical event mapping projects mentioned above. They have researched existing temporal models and timeline visualisations and carried out preliminary scoping of a new spatial-temporal-relationship model for historical events and narrative during Mostern’s Visiting Fellowship at the University of Sydney (July-August 2006).
  • Robertson and Johnson have built an online database of events (crimes, trials, convictions, news reports) for Harlem 1915-1930 (for ARC DP0343148), which breaks new ground in complex database modelling of historical events and on-demand mapping of search results;
  • CoCo is a leader in the development of collaborative learning environments (Reimann 2005) and in science learning by modelling (Reimann & Thompson 2009);
  • Macquarie Library has collaborated with Johnson since 2002 in developing methods of attaching location to database content and representing it online in interactive TimeMap maps, delivered as a component of MacquarieNet, as well as in the production of publication quality mapping of cultural data for the Macquarie Atlas of Indigenous Australia.
References

Chainey, S. & Ratcliffe, J. 2005 GIS and Crime Mapping. West Sussex: Chichester.

Conolly, J. (ed) 2006 Geographical Information Systems in Archaeology. Cambridge: Cambridge University Press.

Cunfer, G. 2002 Causes of the Dustbowl in Knowles, A. Past Time, Past Place: GIS for History. 93-104.

Drucker, J. & Nowviskie, B. 2005 Speculative Computing: Aesthetic Provocations in Humanities Computing in Schreibman, S., Siemens, R.G. and Unsworth, J. (eds) Companion to Digital Humanities. Blackwell. 431-447.

Edelson, D.C. 2001 Learning-for-Use: A framework for the design of technology-supported inquiry activities. Journal of Research in Science Teaching, 38(3), 355-385.

Gilbert, J.K. & Boulter, C.J. 1998. Learning Science through models and modelling. In B. J. Fraser & K. G. Tobin (Eds.) International Handbook of Science Education Vol. 2, pp. 53-66. Dordrecht: Kluwer Academic Publishers.

Jensen, M. 2006 Semantic Timeline Tools for History and Criticism. Proceedings ACH/ALLC Paris July 2006.

Johnson, I. 1997 Mapping the fourth dimension: the TimeMap project. in Dingwall, L., et al. (eds) 1999 Archaeology in the Age of the Internet. British Archaeological Reports International Series 750. 21pp.

Johnson, I. 2002 Contextualising Archaeological Information Through Interactive Maps. Internet Archaeology 12.

Johnson, I. & North, M. 1995 Archaeological Applications of GIS. Sydney Univ. Archaeological Methods Series Vol 5.

Knowles, A. & Hillier, A. 2008 Placing History: How Maps, Spatial Data, and GIS are Changing Historical Scholarship. Redlands: ESRI Press.

Knowles, A. (ed) 2005 Emerging Trends in Historical GIS. Historical Geography 33 (thematic issue).

Knowles, A. 2002 Past Time, Past Place: GIS for History. Redlands: ESRI Press. 222pp.

Langren, G. 1992 Time in Geographic Information Systems. London: Taylor and Francis.

McCarty, W. 2005 Humanities Computing. New York : Palgrave Macmillan.

Petras, V., Larson, R. & Buckland, M. 2006 Time Period Directories: A Metadata Infrastructure for Placing Events in Temporal and Geographic Context. Opening Information Horizons: Proceedings ACM/IEEE-CS Joint Conference on Digital Libraries. New York: ACM Press. pp 151-160.

Pustejovsky, J. et al. 2003 TimeML: Robust Specification of Event and Temporal Expressions in Text. IWCS-5, Fifth International Workshop on Computational Semantics.

Reimann, P. 2005 Co-constructing artefacts and knowledge in net-based teams: Implications for the design of collaborative learning environments. in H. L. Chick & J. L. Vincent (eds.), Proceedings of the 29th conference of the international group for the psychology of mathematics education (PME 29). Vol. 1, pp. 53-68. Melbourne: Univ. of Melbourne.

Reimann, P., & Thompson, K. 2009. Ecosystem modelling for environmental education: From stocks and flows to behavior and interactions. In P. Blumschein, J. Stroebel, W. Hung & D. Jonassen P. Blumschein, J. Stroebel, W. Hung & D. Jonassen P. Blumschein, J. Stroebel, W. Hung & D. Jonassen P. Blumschein, J. Stroebel, W. Hung & D. Jonassen P. Blumschein, J. Stroebel, W. Hung & D. Jonassen (ed.), Model-based approaches to learning: Using systems models and simulations to improve understanding and problem solving in complex domains Rotterdam, Netherlands: Sense Publishers.

Robinson, B. 2004. Introduction to HEML