Enrique Cerrillo Cuenca

Researcher, Institute of History, Centre for Human and Social Sciences

Spanish Council for Scientific Research

The popularization of GIS represented for many archaeologists a powerful mechanism for accessing the most intricate aspects of past landscapes, in what can be considered a milestone in spatial and landscape Archaeology. Today, the novelties come from the field of remote sensing, where the development of semi-automatic detection of archaeological sites seems to be a groundbreaking approach. Through techniques from the field of computer vision and machine learning is possible to identify areas of archaeological interest among hundreds or even thousands of spatial datasets. This procedure has been named as “semi-automatic detection”, considering that the ultimate validation of detected areas is only possible through a careful revision by a specialist. A growing bibliography can show the interest of these approaches among archaeologists, but also their potential risks on the archaeological interpretation of landscape.

In the present paper, we discuss the impact of these technique on the archaeological knowledge of the Late Prehistory Iberia. An artificial neural network (ANN) has been trained to locate topographical features that match funerary mounds. The challenge is to understand if the current distribution of monuments -mainly build upon the data of the non-systematic archaeological activity from 20th century- can be balanced by an unsupervised detection. We will discuss: 1) the theoretical issues behind the design of the methodology, 2) the design and training of an ANN, 3) the preliminary and expected outcomes and 4) the risks of creating new biases in the data. We are also interested in the quality of these type of data and its possible impact on the creation of new knowledge, a problem that is common to other fields of Spatial Humanities.

Pau de Soto (Institut d’Estudis Catalans)

The configuration of a territory is a long-term evolution in which many factors influenced its shape and morphology. One of the most visible elements that help historians to analyse its configuration is the location of the urban settlements and all the transport infrastructures (by land, river and sea) that connected them. In this paper, the Iberian Peninsula transport networks have been analysed in four different timeframes (Roman Times, Medieval Era, Modern Era and the XIXth Century). By comparing the historical transport networks, it is possible to discover the continuities and differences between each period. Those changes let us relate each network with the political and economic situation of the Iberian Peninsula. This information is extremely useful to define the role of the political decisions in the design and organization of this territory and their impact in its economic situation.

After almost 2 years of an intense digitization process, we have obtained high detailed historical transport networks from each time period studied. Methodologically, Networks Science analyses have been applied to the historical intermodal transport systems. The results of these mathematical processes let us visualize and understand the morphological configuration of the transport networks and determine distribution and mobility patterns and their connectivity degree, and the changes and continuities between periods. In this paper, the methodology of the project will also be presented, from the creation of a transport model of time and cost expenses using historical information to the application of weighted SNA to analyse the connectivity of each historical networks.

Ian Gregory and Laura Paterson

Analysing large textual corpora in ways that combine an understanding of the broad patterns within the text, while also preserving the detail and nuance is one of the major challenges facing the digital humanities. Geography represents one way of doing this: the challenge is to understand what is being said about particular places and how this varies from place to place. To do this we bring together methods from corpus linguistics, that allow us to summarise and explore large volumes of text, with geographical information science (GISc), that provides the tools to allow us to explore sources geographically. We term his approach Geographical Text Analysis (GTA).

Quantitative geographies of poverty are well known, however the geographies of the perception of poverty are far less well understood. To explore these questions we make use of two large British newspaper corpora from the period 2010-15. The Guardian gives us the opinions of left-wing broadsheet newspaper, while the Daily Mail provides a contrast from a right-wing tabloid. Together these two newspapers provide us with 380 million words of text from the period. The questions that we explore are: a) what places within the UK do these newspapers associate with poverty; b) what aspects of poverty are they associating with these places; c) how does this vary from place to place, and; d) how the two newspapers differ in their coverage.

To answer these we first use corpus linguistics-based approaches to identify a set of search-terms associated with poverty. Having done this, we then use an approach called concordance geoparsing to identify the place-names that occur near to the search terms. These place-names are allocated to a coordinate. Once the geoparsing has taken place, the overall patterns of poverty in the two newspapers can be mapped. From here, a range of GTA techniques, derived from combining approaches from corpus linguistics and GISc can be used to explore these patterns in more detail and to compare them with each other. We will presents the results of this analysis, while also exploring the wider implications of the advantages and limitations of being able to explore the geographies in large corpora in this way.

Leif Isaksen, University of Exeter

Elton Barker, the Open University

Rebecca Kahn, Humboldt Institute for the Internet and Society (HIIG)

Valeria Vitale, Institute of Classical Studies, School of Advanced Study, London

Rainer Simon, AIT: Austrian Institute of Technology

Since 2011 the Pelagios project has been dedicated to horizon-scanning, mapping and bridging the semantic annotation research landscape in the spatial humanities. It has done so by organising research events, developing software, formulating standards, writing documentation, and supporting new initiatives through an annual programme of small grants. Its community includes content providers, gazetteer maintainers and educators, in addition to more than 3000 registered users of its Recogito annotation platform. Its members vary enormously in the scale of their activities, their motivations and restrictions, their levels of technical literacy, and the domain of their interests across space, time and disciplinary agendas. But at the heart of this endeavour lies a paradox: how can initiatives like Pelagios whose funders expect them to offer unifying guidance and leadership also support diversity and difference? In attempting to classify and coordinate digital ecosystems do we risk constricting or even killing them?

This presentation does not (and could not) attempt provide a solution to this challenge but describes recent and current developments within the Pelagios initiative that are intended to help address it. Foremost among these is establishment of Pelagios as an (unincorporated) voluntary network association guided by a series of six chartered Activities. By clearly specifying the scope of its mission Pelagios is now able to devolve more control to independently managed projects that collectively realise its objectives. In return, these projects are supported in aligning their activities with those of others that can maximise their impact and benefit from their implementation.

Three of the Association’s Activities are dedicated to the theoretical development and concrete implementation of the production, registration and use of semantic annotations. Three more will support gazetteer production and alignment, education & training, and a small grants programme. Each of these activities is led by two people appointed bi-annually by their relevant stakeholders’ community, and collectively they will undertake to further a vision of linked open geodata founded on the principles of collaboration, mutual benefit and reducing barriers to entry for data providers and end users alike. The Association is thus open to any individual engaged within the field to help forge its future direction, and we openly invite expressions of interest to join it. With the Association’s establishment in early 2019, DH2019 is an ideal opportunity to reflect on the early successes and challenges of this new organisational model as means for balancing diversity and synergy in the Spatial humanities landscape.

Tiago Gil

University of Brasilia

Maps have long been used to represent historical knowledge. Lately, this use has been widely diffused, with the publishing of several new works which emphasize cartography as a language for historians. The approaches are diverse and the research problems addressed to these methodologies are numerous. The purpose of this text is to highlight a specific type of cartographic production in history: the use of narrative maps.

Under the expression "narrative maps" there are at least two possibilities of approach. The first concerns the mapping of narratives, that is, converting texts to maps, something close to Franco Moretti's work called "Atlas of the European Novel". The possibilities within this perspective are broad, allowing both the spatial representation of a specific work and that of a single variable in a set of works. In this case, the texts are read and, through various procedures, decomposed in the form of maps, from certain variables chosen by the one who analyzes. It is possible to "map" the course of a certain character in a given source, for instance.

The second approach concerns a different purpose: the representation of the results of the historians' researches in the form of maps oriented simultaneously in time and space. Although this visualization is commonly associated with economic data, it can also be used to describe and analyze other processes, such as population displacements, political movements, pilgrimages, conflicts, circulation in urban spaces, among others. The paper presents the first steps in this field to indicate the innumerable challenges of cartography applied to historical narrative.

Some reference works on the subject have to be highlighted, especially the works of Jacques Bertin and Edward Tufte. Bertin was a geographer and the creator of the “Semiology of graphics”, through which he sought to create "useful" and efficient visualizations of data resulting from research. Tufte, a designer with a background in statistics, sought throughout his career to present models of visual representation that came out of what he called “flatland”, the flat information landscape. His goal had always been the creation of representations that combined multiple variables of analysis with clarity and beauty.

There are at least six two-dimensional forms of cartographic representation of narratives: a) series of images (like cinema), b) sequential arts (like comics), c) vector animations, d) small multiples, e) anamorphosis, f) traditional static maps, provided they have features that represent the narrative, usually through arrows, lines or other elements that indicate spatio-temporal dynamics.

A conceptual apparatus taken from statistics may be important in understanding the visual processes of narrative mapping. Considering that cartography is a language, we can understand that narrative maps have their own dialect. There are several non-traditional elements to take into account and these elements are quite important: rhythm, speed, linearity, continuity and interpolation. In addition, traditional elements such as colors and shapes acquire a different meaning, a particular accent. These elements may allow for clearer communication in the production of maps, especially animated maps.

Dr Andrew Richardson.

University of Northumbria, Newcastle.

The field of digital humanities has seen a growth in the usage and range of digital tools and methods for the visualisation of literary texts. Increased availability of creative digital processes, tools and environments (e.g. games engines and open-source programming environments) have opened up opportunities for the creative exploration of new methods, techniques and approaches for literary visualisation within geo-spatial contexts. It has also encouraged the development of new forms of engagement with literary material for academic and non-academic audiences.

Using practical investigations as sample case studies, this paper will outline the visual possibilities of creative coding and computational methods for the development of novel (‘experiential’) geolocated representations of literary texts. It will examine the opportunities afforded by computational processes for applying the spatial qualities of locational data to the visual representation of the words; outline methods, applications and outcomes applied to specific examples of literary texts; and provide a platform for discussion into current and future roles of creative digital methods and technologies in this field.

Project examples will be used to highlight individual investigations into the use of locational data and creative technologies to produce new types of ‘experiential’ textual encounters in both virtual and physical spatial environments. Applications of Augmented Reality (AR) technology for mapping and representation of the works of James Hogg in the physical landscape of the Scottish Borders will be used as a basis to discuss ideas around the representation of texts within real spaces and locations. The use of GIS and GPS data to generate geo-spatial visualisations of William Wordsworth’s The Prelude as a textual ‘landscape’ of real and imaged pathways will highlight ideas around the use of spatial data as a ‘ground’ for generating real and imagined literary landscapes.

The presentation of digital and interactive visual outcomes from these practical projects will demonstrate some of the possibilities, opportunities and challenges of computational and game technology for (re)locating texts in digital and physical spaces; outline different types of ‘experiential’ encounters with text as landscape and text in landscape; and raise questions around the possibilities of using physical spaces as a ground to help generate new ‘fictional’ textual encounters and landscapes. The work contributes to a broader discussion around the use of computational methods to develop novel geo-spatial techniques and their possible application in wider contexts. The discussion of processes and outcomes in this presentation will be used as a basis to outline current approaches and discuss future possibilities for the application of creative digital technology towards new spatial visualisations and encounters of text in academic and public-facing contexts.