Dynamic ontology

Advances in archaeological urban theory emphasise dynamic, bottom-up processes (Raja & Sindbæk Reference Raja and Sindbæk2020; Leadbetter Reference Leadbetter2021; Smith Reference Smith2023) and this is reflected in URBank’s data model. Each city is represented as the centre of a web of connected elements—for example, infrastructure, land use and chronological data (Figure 1). In network terms, this is a graph: a set of points (nodes) and their connections (edges). This design accommodates urban complexity and allows researchers to interrogate the very concept of ‘the city’ by analysing the relations from which it arises. It also enables operationalisation of diverse city definitions, facilitating comparative analyses without re-collection of data.

Figure 1. URBank Data Concept: cities are represented as abstract nodes at the centre of a graph with other nodes representing varied associated spatial and temporal data (left panel). This graph representation can be translated into GIS data layers for spatial analyses (right panel) (figure by Hans Sell).

URBank is a platform for addressing fundamental questions about ancient cities. Many of the most important advances in urban archaeology, from identifying scaling laws (Ortman & Coffey Reference Ortman and Coffey2017) to comparing political economies across regions (Drennan & Peterson Reference Drennan, Peterson and Smith2012), derive from the compilation of small, often messy datasets. Important, theoretical and policy-relevant questions—such as, do ancient cities exhibit agglomeration economies comparable to those in modern systems? Or, how do population densities and spatial layouts differ between urban traditions, and what do those differences reveal about their political and economic organisation?—can be tackled rigorously with the kind of systematic aggregation URBank enables. By establishing common standards for metadata and coding, URBank turns scattered, one-off datasets into a versatile, cumulative resource.

Open FAIR science

URBank is committed to Open Science and FAIR principles (Wilkinson et al. Reference Wilkinson2016) (Figure 2). The database schema and metadata are published in machine-readable formats, and records have persistent, unique identifiers, including a digital object identifier (DOI) and universal urban centre identifier (UUCID) (Figure 3). Versioned snapshots of the database are archived upon the finalisation of new entries.

Figure 2. The FAIR principles upon which URBank is built (figure by Hans Sell).

Figure 3. URBank’s UUCID logic (figure by Hans Sell).

URBank also recognises the complicated ethical and legal dimensions surrounding archaeological data. While all data will be accessible to verified partners and institutions, public access is governed by a tiered system. Contributors may mark entries with different levels of access and co-ordinate precision, ensuring compliance with national regulations and community priorities. In cases where restrictions prevent full sharing, URBank maintains UUCIDs as reference records and directs users to responsible institutions or agencies.

Designed to evolve, built to last

URBank is hosted at the Max Planck Institute of Geoanthropology (Jena, Germany) and developed in collaboration with leading urban archaeologists, historians, urban studies scholars and Earth System scientists (https://urbank.earth). Launched in January 2025 by the Department of Coevolution of Land Use and Urbanisation, and founded with partners including Durham University and the University of Colorado Boulder, URBank’s data ontology is already a fundamentally new resource for comparative urban research. It is, however, explicitly designed to evolve and to respond dynamically to new ideas, collaborations and theoretical advances. It serves as a global hub for aggregating and analysing urban data, supported by the Coalition for Archaeological Synthesis (CfAS) and dialogue with international scholars. With hundreds of entries already, URBank is actively being populated by the CfAS Urban Adaptation Working Group (https://www.archsynth.org/project/urban-adaptation-to-environmental-change/).

URBank in action

URBank’s advantages are highlighted by two examples. In the first, a low-dimensional dataset of pre-Classical urban centres from South-west Asia and Egypt are represented in the traditional cities-as-points format (Lawrence et al. Reference Lawrence, Hinojosa-Balino and Chelazzi2025) (Figure 4A). Points represent central co-ordinates for settlement mounds, and each location links to a corresponding city construct in the database graph with attributes drawn from spreadsheet columns.

Figure 4. URBank supports a spectrum of data types and visualisations, from ‘city-as-point’ data (panel A, showing pre-Classical urban centres from South-west Asia and Egypt) to ‘city-as-constellation’ data (panel B, showing Koh Ker, Cambodia) (figure by Hans Sell).

In the second example, rich, multilayered datasets from Koh Ker, Cambodia, briefly the capital of the Angkorian World during the tenth century (Tho et al. Reference Tho, Huon and Moul2016), highlight how URBank can also represent relational urbanism. Koh Ker is a constellation of temples, houses, water infrastructure and land use (Figure 4B). Over time, as the graph of interlinked records grows, it will become possible to evaluate whether constructs like Koh Ker have empirical coherence—whether they correspond to temporally persistent patterns in the lower-level graph of data—indicating that a single, persistent ‘city’ actually existed.

URBank enables the use of both simple ‘city-as-point’ and detailed ‘city-as-constellation’ modalities. The quality of the datasets, rather than database architecture, shapes analytical possibilities.

Outlook

URBank consists of a standalone database and services that integrate partner data, with a layered architecture reflecting this dual purpose (Figure 5). This structure will allow URBank to be a primary analytical resource for the study of global urbanism across human history. The project welcomes contributions and collaborations, developing data structures and theoretical approaches that connect archaeology to contemporary challenges. Researchers interested in linking their urban data to URBank can request an input template from the corresponding author.

Figure 5. URBank’s multilayered architecture (figure by Hans Sell).