University of Southampton OCS (beta), CAA 2012

Font Size: 
Linking with legacy: Modelling spatio-temporal distribution patterns of 40 year old excavation data from the settlement site of Zagora
Ian Johnson, Beatrice McLoughlin, Matthew McCallum, Andrew Wilson, Steven Hayes

Last modified: 2011-12-16


Excavations at the Greek Early Iron Age settlement of Zagora (Andros) were carried out by an Australian team in the 1960s and 70s. Efforts are now well underway to analyse and publish the complex manually collected legacy data, using Heurist a flexible collaborative eResearch database designed for Humanities data.

Although it remains the most extensive preserved town plan of a 10-8th century BC site known in the Aegean, the site’s full value as a window into a period of rapidly changing social organization between the collapse of the Mycenaean palace economies and the beginning of early modern state polities has until now not been possible.

The Australian excavations uncovered 10% of the site, encompassing at least 25 households, large and small, clustered in discernible neighbourhoods. The excavation records include 200 architectural plans, 6000 site photographs, more than 1000 pages of field notes, 3700 inventory records  and their drawings and photographs. The houses preserve a substantial amount of domestic cooking and storage pottery, imported drinking and serving wares, everyday utensils such as weaving equipment, and the faunal remains representing dietary practice.

While preliminary reports of the excavation were published at the time, it was not possible for the excavators to visualise [model] the nuanced patterns of the material culture and what it could tell us about similarities and differences between households and neighbourhoods due to the complex interplay of relationships involved in understanding this material. Today we can combine database technology and GIS for both analysis and presentation of the material.

The database needs to represent the relationships between these categories of data. It needs to be web-based to support a distributed network of researchers, and provide ways of building and annotating interpretive links between entities in the database as the analysis of the data proceeds. It needs to be able to evolve with increasing understanding of the data and its structure as well as allow for incorporation of data from future excavations - the design is unlikely to remain static.


A web-based database of the data has been built using Heurist ( a research database tool developed at the University of Sydney Arts eResearch unit. Implementation of the Zagora project database has run in parallel with the development of Heurist version 3 allowing close cooperation between the project archaeologists and the programming team (who coincidentally are housed a few doors apart). This has allowed refinement of the Heurist database model to more easily support complex archaeological data.


Zagora provides a valuable case-study of a long-term excavation combining large amounts of legacy data with an ongoing research program. In this paper we will focus on the specific features of Heurist which are of particular assistance to the Zagora project. These include:

  • the enforcement of structural relationships between entities;

  • the ability to build interpretation by linking entities through annotation;

  • sharing data while retaining privacy of notes and interpretation;

  • selective publication of information, including interpretations, to a project web site;

  • live integration of the database with ArcGIS;

  • use of the database in the field.

We will illustrate the modelling of the Zagora data and reflect on the advantages conferred by the flexibility of a generic database over the costs and restrictions imposed by development or adoption of a bespoke system. We will conclude with a discussion of the importance of exposing data and interpretation as part of the publication strategy rather than limiting publication to traditional linear accounts (monographs and papers).


excavation, database modeling, GIS, legacy data