quarta-feira, 23 de outubro de 2013

Networks over space and time (abstracts)

Abstracts

Pau de Soto Cañamares (Instituto de Arqueología de Mérida - Spain): Costs & times of the Roman transport. Using network analysis to understand the Roman transportation system
Several methodological approaches are used to these days that also suggest how the Roman transport worked. This project is based on the operability of the Roman infrastructures as an indispensable way to know the benefits and shortcomings of the transportation system created in Roman times. A thorough analysis of each distribution models set (both temporary and costs) provides valuable information for understanding the mechanisms of the Roman economy and society. It is therefore obvious that the combination of all of the approaches (archaeological material, ancient sources, network simulation...) should allow us to obtain a more global perspective of the Roman economy, especially in matters of movement of goods.
The main geographical focus of this project is the NE of Hispania, but with the aim of use these methodology in a much broader geographic frame, the entire Iberian Peninsula, Italy and Britain were analyzed.
As would be seen during the presentation of this work, the knowledge of infrastructures is essential to obtain a more accurate knowledge of the freight transportation. This project has taken into account while analyzing the whole infrastructure of Roman roads which existed in Roman times, whether through land environment, river or sea. A set of constant values have been used to calculate the costs and transportation time needed for commerce. So, this model offers a simulation of possible costs and times needed to transport certain goods that had to be spent to travel from a particular spot of territory to another (and even the entire network).
Finally, the ability to see graphically and quantified those costs and time values which until now they could only be guessed, can open new perspectives and justifications to the speeches made on the work done until today. In fact, the comparison between these results and the analysis of archaeological and historical interpretations should not invalidate the final information but in many cases they should complement each other, clarifying and offering more elements for a global vision.


Thomas Thevenin (University of Burgundy - France), Robert S. Schwartz (Mount Holoyke College - USA) and Christophe Mimeur (University of Burgundy - France): Measuring the link between space and network over time
The railway growth seems to be essential to the economic dynamism of the north in France from 1830 to 1930. However, could we underline the same statement for the south of the country and for general consideration for the rural areas? On one hand, the transportation economy the network effect is essential to develop economy, agriculture and demography. Many governments’ policies are based on this mythic belief to justify the construction of important infrastructures. On the other hand, many others authors as historians and geographers criticized this position on the “network effect” (Pumain 1982). These works are usually based on aggregate scale or are focused on urban areas. The database presented in this article could be used to work at different scales to consider rural and urban regions or agrarian or industrial sectors on a long period of time.  In this way, we need to explore the explicative power of econometric solutions. We will present the first encouraging results based on GWR indicators (Fotheringham, Brunsdon, and Charlton 2009). This measure will be essential to pass from a descriptive approach to an explicative scientific strategy. 

Albertina Ferreira (Instituto Politécnico de Santarém - Portugal), Carlos Caldeira (Universidade de Évora - Portugal) and Fernanda Olival (Universidade de Évora - Portugal): From low density networks to geo-temporal approach
This study is based in 117000 prosopographic registers available in the SPARES database (Prosophographic System of Social Relations and Events Analysis). This database has being developed by the research project “Intermediate groups in the Portuguese dominions: the ‘familiares’ of the Holy Office (c. 1570 -1773)”, at the University of Évora. The database collects information regarding biographic and relational events, from the sixteenth to the eighteenth century. All the data is geo referenced. When producing historical maps about the location of familiares and comissários of Inquisition (1575-1775), the research teams has realized the existence of large areas of low density distribution of these characters’ networks. This study aims the creation of an analytic geo-temporal model which would allow historians to study these areas of low density distribution, in a comparative way. Departing from the dynamic networks analysis approach, this methodology tries to adapt it to the elaboration of new research parameters. In this sense, the team tried to coordinate the database with geographical information system software, the ArcGIS. Even if these trials will be constructed from the Inquisition historical data, these model can be applied to the other research themes involving time/space/networks.

Martin Stark, (University of Hamburg - Germany): Locating historical networks in time and space: current achievements and challenges
Given the nature of their sources, network analysis approaches in history often have a primary focus on social interactions. Well-known examples are letter exchanges between scholars, traders, covert resistance activists, credit markets, career advances or migrations. Many of the above mentioned network studies have however strong spatial components as well, which directly affect the creation of social ties, their maintenance and nature. Medieval trading for example depended largely on the capacity or failure to cover geographical distances, while the speed and intensity of scholarly exchanges depended on the reliability and speed of postal systems. Research in rural credit markets has for example revealed strong cross-border ties between 19th century Germany, France and Luxembourg. At this stage it seems however that by and large the spatial dimension acts as a background against which historical social interactions are being studied. I will present a selection of case studies and their integration and exploration of the spatial dimension.
Historical sources not only allow us to reconstruct social interaction in detail but also offer clues with regard to temporal dimension in which they occurred: Serial sources such as church registers, trading contracts and letters are often very easy to date and relate to each other in time. Network analyses which are based on the hermeneutic analysis of texts and other objects typically need to deal with heterogeneous data: Some ties can be dated precisely to an hour, whereas in other cases scholars need to infer time stamps based on the context of other events or simply can not make any such statements at all. I will discuss the challenges posed by missing data and data collection methods as well as the challenges inherent in exploring temporal data using different visualisation techniques, some generic, some tailored to the needs of specific research questions.

Tim Evans (Imperial College London - UK): Spatial Network Models in Archaeology
I will look at the spatial network models that have been used in archaeology. I aim to show to what extent they are all part of large families of models which will highlight the similarities and the differences. I will then ask if one model is better than another and how we might answer that question. I will also look at the sort of questions that can be answered with such models.

Joaquim de Carvalho (Universidade de Coimbra - Portugal): Networks, self-organisation and historical research: uncovering hidden structures in historical data
The theme of self-organisation and the emergence of complex structures has been object of intense interdisciplinary interest since the beginning of the century. Historical research has been somewhat distant from these new approaches, certainly because of methodological and empirical difficulties in finding opportunities of applying such concepts to concrete historiographical problems supported by historical sources. We will demonstrate that it is possible to detect historical processes in which there is strong evidence of mechanisms of self-organization at work. We also show how common sources contain precious information that can be made visible by applying special network analysis tools. We will focus on two examples: the choice of godfathers as recorded in parish registers and the circulation of mail in the 18th century. The main conclusion from our examples is that historians should bring into their conceptual and methodological tools the findings of Complexity Science, namely the concepts of Emergence and Self-Organization, and the techniques of network reconstitution and analysis. By incorporating tools and concepts such as these new insights can be gained into the fundamental questions of the persistence of structures and the interaction of structures and individual agency.

Clement Levallois (Erasmus University Rotterdam - Netherlands): Visualization of large and time-dependent networks: advances and limits
Network visualizations are helpful devices for the exploratory analysis of a dataset and are increasingly accepted as legitimate formats for the visual display of an argument in the social sciences and the humanities.
I will report on recent advances in software development (evolutions of the Gephi platform) which widen the scope of these visualizations: the acceptable size of datasets becomes larger, and datasets of such a large size and with a time dimension can be represented.
Experimenting with these new possibilities opens the question of the meaning of the visualization thus performed. Based on the visualization of a large dataset of transactional data, I will discuss how (still young!) conventions for the meanings attached to the visualization of dynamic networks are challenged by the scale and transactional nature of the dataset.
These advances are themselves anything but stabilized results, and the conclusion will discuss questions that are opened in the representation of large, time-dependent networks.

Sofia Oliveira, Jared Hawkey and Nuno Correia (CADA and Universidade Nova de Lisboa - Portugal): Finding and Representing Personal Time/Space Patterns
The talk describes the work carried out in a project, Time Machine, that aims to stimulate reflection about personal routine while engaging in a dialog regarding the daily uses of ubiquitous computing, and a more broad discussion regarding the methods and relations between art and science. TimeMachine was proposed as a collaborative project between CADA, a Lisbon-based art group that creates playful-experimental software mainly using mobile technologies, and the Interactive Multimedia Group of CITI/FCT/UNL, that works on different aspects of describing, processing and interacting with multimedia information. One of the main outcomes of TimeMachine is a mobile application that captures and processes location data and creates personal and intimate time and space maps that capture routine and activity. Visual representations exploit color, shape and proximity to show the network of meaningful places and how they are organized temporally. The visual representations rely on  a carefully designed and rigorous processing framework that enables concrete representations of time and space but also supports the development of subtle and ambiguous representation forms. The work was developed in an iterative process where multiple processing and visualization prototypes were developed, tested and subject to critical reviews. The talk will discuss the different methods that were employed to develop the project, the results obtained so far and open issues for further research. Particular attention is dedicated to the tension created by the different goals that the project had considering its desired artistic and scientific outcomes. http://img.di.fct.unl.pt http://cada1.net

quarta-feira, 9 de outubro de 2013

Networks over space and time: modelling, analyzing, and representing complex data in the digital humanities


November 8th 2013, FCSH, New University of Lisbon, Lisbon, Portugal

WG1, Spatial and temporal modelling: representation of space and time, NeDiMAH

(Auditorium 2, Tower B, 3rd floor)

The workshop is free. All those who wish to attend please register here:

More info (programme - travel grants)

This workshop is about interconnections between, and in space and time. But it also sees interconnections at other levels: between modelling and analysing, between theory and practice, as well as between humanities and computing.

In the humanities, a close look at networks and relationships, whether formal or informal, personal or social, of information or of knowledge, of transportation or of communication, has always been an important subject of study and, at the same time, a powerful analytical process. In computer science, the study of networks and of methodologies for analysis and visualization of these relationships is nowadays an increasingly well understood and practiced area of knowledge. In both the humanities and computer science, researchers are well aware of the dynamic nature of data and knowledge when viewed through the lenses of space and time.

Networks can be studied in a purely spatial perspective, if the object of analysis is the distance between things or people. However, there are two other dimensions which render networks’ study in a more complex and richer methodology. Either time or social relationships help to extend the focus of analysis from distance to connectivity, and this is an important concept for the Humanities, as it is for the Social Sciences, at least, since the 1930’s . In the field of spatial analysis, the focus has also tended to shift from an almost exclusively quantitative approach , to one that tries to develop a new ontological and epistemological view, combining quantitative with qualitative methods and sources , a view also important for the humanists. When put together, time, spatial analysis, with its derivative, spatial network analysis, and social network analysis, can be a powerful way of thinking about the world (theory) and of explaining it (methodology). And at the present time, with the integration and plasticity of the digital, the rising awareness about geography and time trough the Internet’s social networks, and the growing usability of the Web 2.0, thinking and explaining networks can benefit from powerful tools, increasingly complex and accessible at the same time.

The aim of this workshop is to combine analytical perspectives in the study of networks, over space and time, in humanities disciplines and on various themes, to identify methodologies, discuss research results, and encourage interdisciplinary approaches. The main focus of this workshop will be the areas of modelling and representation, highlighting them more as methods of analysis and knowledge production than merely as tools.

The perceived outcomes of this workshop will be to document the case studies presented, discuss and share methods and research results arising from these and to identify in the form of a report the interconnection between the humanities and the digital, helping to define a taxonomy of new methodologies and the development of a community of researchers for future collaborative work.