In this workshop research data becomes the proverbial rubber that meets the road. Models, data and interpretations from a range of research domains will be juxtaposed and illustrated by worked case studies. This will include data sampled from a present population in a controlled manner, as well as more extreme data about past populations that is fragmentary and results from only weakly controlled sampling. An overarching aim will be to identify the manifold sources of bias, uncertainty and error in real-world research data, and to scrutinize data-derived conclusions. Another focus will be on formal methods and frameworks for handling incomplete data and representing uncertainty. The workshop is part of the Thematic Einstein Semester “The Mathematics of Complex Social Systems: Past, Present, and Future”.
Deadline for registration: May 5, 2022. Registration is free of charge, but mandatory.
To register please use the link: https://www.conftool.net/tes-summer-2022/
We welcome submissions for posters on the topic of Workshop II “Data Past and Present”. Submissions should summarize the work to be presented on max. one pdf page including a descriptive figure and references. The PDF file with your abstract can be uploaded in the registration platform for the event by April 30, 2022.
Laura Alessandretti (Technical University of Denmark), Ferran Antolín (German Archaeological Institute), Michael Barton (Arizona State University), Andrew Bevan (University College London), Tom Brughmans (Aarhus University), Christine Hertler (HADW Senckenberg Forschungsinstitut), Brigitta Schütt (Free University Berlin), Philipp Lorenz-Spreen (Max Planck Institute for Human Development), Michael Mäs (Karlsruhe Institute of Technology), Camille Roth (CNRS and HU Berlin), Philip Verhagen (Vrije Universiteit Amsterdam)
The workshop takes place as a hybrid event, where the on-site part will be at Zuse Institute Berlin in Berlin-Dahlem.
|9:00 – 9:30||
Opening (Nataša Djurdjevac Conrad, Benjamin Ducke)
Michael Mäs (Karlsruhe Institute of Technology): Digital-trace data: the end or the beginning of theory?
Laura Alessandretti (Technical University of Denmark): Modelling individual mobility behaviour
|13:30-14:30||Camille Roth (CNRS and HU Berlin): Echo chambers, fragmentation of digital public spaces, filter bubbles? It depends on data.
|14:30-15:30||Philipp Lorenz-Spreen (Max Planck Institute for Human Development): Long-term data and Field Experiments on social media|
|16:00-16:30||Ferran Antolín (German Archaeological Institute), Maria Elena Castiello (University of Bern): Human and agricultural niches in the western Mediterranean area during the Neolithic: from data to models
|16:30-17:00||Brigitta Schütt (Free University Berlin): The Transformation of the Pergamon Micro-Region between Hellenism and Roman Imperial Period
|17:00-17:30||Nataša Djurdjevac Conrad, Benjamin Ducke: MATH+ Project Overview|
|from 17:30||Poster Session and Get together
|9:00-10:00||Christine Hertler (HADW Senckenberg Forschungsinstitut): Hominin dispersal across scales – An example for the integration of agent-based modeling|
|10:00-11:00||Andrew Bevan (University College London): Adversarial Approaches to Data, Methods and Models in Archaeology|
|11:30-12:30||Tom Brughmans (Aarhus University): MINERVA: ceramic data analysis and Roman road data collection|
|14:00-15:00||Michael Barton (Arizona State University): Regional Data and Regional Analyses for Regional Systems|
Laura Alessandretti (Technical University of Denmark): Modelling individual mobility behaviour
Abstract: From choosing a restaurant for dinner to deciding how to get there, spatial decisions are ubiquitous in human day-to-day lives. Taken together, these choices underlie critical societal phenomena, including the spread of epidemics, the emergence of traffic congestion, and urban segregation. In this talk, I will present recent research that leverages high-resolution large-scale mobility data to understand some of the mechanisms underlying individual mobility. The talk will touch upon key aspects, such as the interplay between exploration and exploitation, the effect of cognitive constraints, the relation between social and spatial behavior, the effect of spatial scales, and differences across genders and age-groups.
Ferran Antolín (German Archaeological Institute), Maria Elena Castiello (University of Bern): “Human and agricultural niches in the western Mediterranean area during the Neolithic: from data to models”
Abstract: Neolithic farmers arrived in the Western Mediterranean area from the East, established their settlements in coastal areas and, over time, migrated to new environments, showing capacity to adapt to changing ecological and climatic conditions. Farming practices and the nature of settlements differ greatly from those known from the Eastern Mediterranean and central Europe. To which extent are these differences connected to the local environment? The AgriChange project (funded by the Swiss National Science Foundation, 2018-2022), with support from the GroundCheck Programme at the DAI, is tackling this question by integrating multiple proxies at a supraregional level, including archaeobotanical data, radiocarbon dates and palaeoclimatic models. These datasets have allowed to observe the pace of spread of farming in the area, the main agricultural and land use dynamics and to locate densely settled areas over time. The currently on-going research is focusing on a multi-proxy and machine learning approach to investigate the impact of ecological and climate constraints on the first Neolithic niches of humans and crops. We modelled and estimated the potential areas suitable for settlement location and for discriminating distinct types of crop cultivation, under changing climate scenarios that characterize the period cal. 5900 – 2300 BP. The results of this study have helped characterize not only the past climate variability and its influence on settlements distribution and crop adaptation in the Western Mediterranean area, but also to pointing out sensitive parameters for a successful application of machine learning procedure, from continental to regional scale.
Michael Barton (Arizona State University): Regional Data and Regional Analyses for Regional Systems
Abstract: Archaeological data come traditionally and most frequently from the excavation of individual locales (i.e., “sites”), where even the most extensive and long running projects expose a few hundred square meters at most. Yet human societies operate at much larger, regional scales than the spatially tiny windows of sites. Foragers have ranges that extend far beyond individual camps. Villages are surrounded by fields and pasture lands, and utilize the hinterlands beyond. Towns and cities encompass activities at even larger scales, with intra- and inter-regional exchange, and distant resource acquisition. I will present approaches to synthesizing and analyzing data at regional scales, using examples from research in the western Mediterranean as examples. I will also discuss some of the challenges of regional-scale research and ways to overcome them. Examples will range from synthesizing datasets from multiple projects carried out by different researchers, to analyses of time-space dynamics of surface collections, to empirical validation of regional scale computational models. For students who would like to experiment with these approaches, most of the examples to be presented were generated with R and GRASS GIS.
Tom Brughmans (Aarhus University): MINERVA: ceramic data analysis and Roman road data collection
Abstract: The goal of the MINERVA project (funded by the Independent Research Fund Denmark as a Sapere Aude research leadership grant) is to enhance our understanding of the Roman economy’s long-term functioning through three lines of research: (1) analysis of long-term ceramic data patterns, (2) high-resolution Roman road model, (3) simulation experiments testing theories explaining ceramics transport and distribution.
1) Aggregation and analysis of large open ceramics datasets, including tableware in the Eastern Mediterranean, and amphorae in the German provinces and in central Italy. Ceramics are copiously available, and well-studied and described in traditional scholarship. Ceramic tableware and amphorae offer comparable data throughout centuries and the entire Empire, allowing for the study of changes in the intensity and direction of inter-regional trade. The statistical analysis of ceramic dataset focuses on changes through time in the geographical distribution and in the volume and proportions of ceramic types at sites, and trends in correlation between distance from place of production with ceramic volume and diversity.
2) Developing a highly detailed open model of the Roman transport network system. Existing digital models of the entire Roman imperial transport system exist at a coarse level of detail that is not representative of our current knowledge of Roman roads or of geographical structuring. We aim to draw on all available historical and archaeological data to develop a model of the Roman road system in high detail across the entire Empire. The digital data will be made openly available on the online gazetteer Itiner-e. It aims to become an online platform for a community of scholars to explore, query, download and edit historical road data, leading to a continually improving resource. This dataset is developed in close cooperation with the project Viator-e (https://viatore.icac.cat).
3) Perform computational simulation experiments to explore whether current theories about the distribution and transportation of foodstuffs and tableware can explain the observed ceramic data patterns. Particular attention is paid to the explanation of centuries-long patterns, with the aim of identifying macroeconomic trends and cycles of inter-regional trade in the Roman economy that only reveal traces over long time periods. Agent-based models informed by transport variables and the highly detailed transport network model will be created to perform these simulation experiments.
Christine Hertler (HADW Senckenberg Forschungsinstitut): Hominin dispersal across scales – An example for the integration of agent-based modeling
Abstract: Early hominin dispersal is examined in paleoanthropology as being empirically reflected by the spatiotemporal distribution of fossil specimens. A fossil is discovered, it is dated and assigned to a particular hominin taxon. The geographical position of the locality proper is then compared to the distribution range of hominin specimens of the same kind, and the chronological sequence of such maps is commonly understood as an expression of underlying dispersal processes. Although the procedure seems to be largely driven by empirical data, the further examination of such processes requires the integration of diverse models on various scales.
Agent-based models are commonly applied in such studies, on a large scale for instance in the examination of large-scale dispersal events like the first dispersal of early hominins out of Africa. In those models, movement is modeled as a diffusion process. Hominin agents are conceived as intrinsically moving and their movements are directed by environmental factors or interactions among the hominins. Moreover, the study of events on smaller scales permits examining the crossing of dispersal barriers on regional scales, for instance potential crossings of sea straits by culturally varying strategies, e.g. swimming, drifting, or rafting.
In this contribution, we will examine the particular role, agent-based models may play in simulations of early hominin dispersal. Agent-based models rely on sets of models for both, the environments, and the agent behavior proper. The integration of data generated by simulations permits to study such events on various scales. We will have a closer look on swimming performances, sea straits and culturally acquired skills in the crossings of sea straits as an example for the integration of modelling approaches in simulation studies.
This work is done together with: Ericson Hölzchen, Jan-Olaf Reschke and Iwan Pramesti Anwar.
Brigitta Schütt (Free University Berlin): The Transformation of the Pergamon Micro-Region between Hellenism and Roman Imperial Period
Abstract: In times of climate change and the discussion about the beginning of a new era shaped by humans (Anthropocene), historical human-environment relations as a research subject are increasingly popular, too. The reconstruction of the complex interactions between nature and civilisation requires interdisciplinary cooperation between humanities, engineering science and natural sciences. Only on this basis our understanding of the old world can be extended by the so far often neglected social-ecological aspects.
One way to achieve this goal is through the study of micro-regions, which are not only geographically defined, but also show common cultural networks and economic cycles. Ancient Pergamon and the surrounding landscape in north-western Turkey provide a good example. For 140 years, archaeologists have been working here together with numerous other disciplines under the aegis of the German Archaeological Institute and with the permission of the Ministry of Culture and Tourism of the Republic of Turkey to explore the metropolis. Its settlement history starts in the 2nd millennium BC, but traces in the surrounding area go back to the 7th millennium. The most substantial and significant evidence so far is available for the Hellenistic Period (3rd-1st century BC) and the Roman Imperial Period (1st-3rd century AD). This particular potential and an interdisciplinary strategy for its scientific evaluation have been awarded by the German Research Foundation (DFG). The project involves archaeology, building research, and physical geography working together with numerous other historical and natural science disciplines.
At the beginning of the project is the question of what interactions existed between profound urban change in Pergamon (marked by a doubling of the urban area since the late 1st century AD and monumental construction) and changes in the micro-region. Based on the observation that the western lower plain of the Kaikos (Bakır Çay) with the adjacent mountain ranges and the coastal zone was characterised first by settlement concentration and demilitarisation since the end of the Pergamenian Royal Period in 133 BC, and later in the Imperial Period by the establishment of leisure or wellness elements such as thermal baths, the relationships between cities, rural settlements and the landscape are examined systematically and diachronically for the first time in the entire micro-region. This involves different levels such as resource use, production and consumption, lifestyle and health of inhabitants, architecture and construction, and the design and perception of living spaces. Given the diversity of natural resources, the proposed project focuses on soil, water, wood, stone, and clay, whose importance to the economic and living environment can be understood only through the direct collaboration of archaeology, building research, and physical geography.
Project leaders: Felix Pirson, German Archaeological Institute, Istanbul Department; Thekla Schulz-Brize, TU Berlin, Historical Building Research; Brigitta Schütt, FU Berlin, Physical Geography. Copy from https://www.dainst.blog/transpergmikro/about-the-project/
Philipp Lorenz-Spreen (Max Planck Institute for Human Development): Long-term data and Field Experiments on social media
Abstract: Computational Social Science has emerged in recent years as an interdisciplinary field, spanning between sociology, political science, but also complexity science and math. This development is partly due to new means of measuring human behavior at a large scale and through passive observation “in the wild”. Even though data access is often restricted, various ways of measuring and quantifying the informational and social behavior of people online have been opened up for research. Researchers recently became interested in combining observational data with experimental methods, in order to be able to assess causal relations. Such field experiments usually link participants’ data from traditional methods, like surveys, with behavioral data, like news consumption or social interactions, recorded via social media or the browser. I will give an overview of the latest developments of digital media field experiments and present a study that makes use of Twitter’s API to get real-time motives for sharing information on Twitter. In addition to field experiments, long-term observational data allows for insights into how behaviour changed on social media since its widespread adoption. I will present a study that shows evidence for social acceleration on Twitter over almost ten years. Overall, my aim is to present how such techniques allow for the collection of high quality data, which can be used for mechanistic insights and modelling and to better understand social media behaviour.
Michael Mäs (Karlsruhe Institute of Technology): Digital-trace data: the end or the beginning of theory?
Abstract: The emerging discipline of Computational Social Science (CSS) is making important contributions to the understanding of human behavior, exploiting novel sources of data from the Internet. On the web, humans leave digital traces that can be measured at low costs and in huge amounts. What is more, most data is observational and does not depend on reactive measurement methods, unlike surveys and laboratory experiments. Inspired by the new opportunities, some experts even proclaim “the end of theory”, arguing that there is so much high-quality data available that theory is no longer needed. While the notion of the end of theory neglects fundamental limits of empirical observation, it also fails to consider the complexity of the social systems that humans build on the web. In this talk, I will demonstrate that neglecting this complexity can lead one to false and potentially very problematic conclusions. I advocate an approach that combines empirical methods from CSS with rigorous formal modeling of complex systems and argue that this approach can support the development and regulation of online social networks.
Camille Roth (CNRS and HU Berlin): Echo chambers, fragmentation of digital public spaces, filter bubbles? It depends on data.
Abstract: Echo chambers refer to the existence of cohesive groups of actors who principally interact with same-minded people. Are they ubiquitous in social media? It might depend on which data one is looking at. At a more macro level, is the notion of online fragmentation, or the co-existence of communities between which there are few bridges, at odds with the idea that online spaces foster global arenas ? It might also depend on how data is filtered. The term of filter bubbles describes the possibility that algorithmic personalization would confine users to consumption of content they are most comfortable with. Is it a necessary behavior of recommendation algorithms? Again, it might depend. We will review these three connected contemporary debates on the social dynamics of online communities, whereby the literature appears to bring conflicting results and where, we argue, the way data are being appraised may play a key role.
Philip Verhagen (Vrije Universiteit Amsterdam): Linking theory, past practice and evidence of ancient movement in computational modelling
Abstract: Since the introduction of GIS in archaeology, there has always been a strong interest in the modelling and reconstructing past movement patterns through the use of least cost path modelling. At first, the mechanics of human movement were the most prominent concern of archaeologists applying LCP models. The least cost principle itself was heavily criticized, and various approaches were developed to enrich archaeological LCP models with cost criteria such as ease of navigation, purpose of travel, and security, allowing us to model a wide variety of movement practices.
It is only fairly recently that these models have become part of a more ambitious agenda to understand ancient movement systems by reconstructing the networks of connections that allowed people and goods to move around. Modelling and reconstructing (parts of) movement networks is beset by a host of problems, mostly due to the difficulty to reconstruct ancient movement practices on the basis of archaeological data, and the fragmentary nature of the evidence for ancient routes and connections, but also due to the increasing computational load involved when scaling up LCP models. Recent studies in LCP modelling are therefore increasingly relying on High Performance Computing to compare and validate reconstructed routes and networks.
In this paper, however, I want to take a step back and focus on the problems related to the nature of the available evidence and its interpretation. In archaeology, establishing the links between theory, ancient practices and material evidence is never easy, and movement is no exception to this. I will demonstrate how we can approach the investigation of ancient movement by presenting a conceptual framework (the ‘track graph’) that we developed as a basic ontological structure for both detecting and modelling ancient movement patterns. Importantly, this framework can allow us to disentangle the archaeological evidence from the processes that gave rise to this evidence, provide the opportunity to test various assumptions on ancient movement in a structured way, and compare case studies. However, implementation of this concept in a simulation modelling context is still a work in progress.
Reference: Nuninger, L.; Verhagen, P.; Libourel, T.; Opitz, R.; Rodier, X.; Laplaige, C.; Fruchart, C.; Leturcq, S.; Levoguer, N. Linking Theories, Past Practices, and Archaeological Remains of Movement through Ontological Reasoning. Information 2020, 11, 338. https://doi.org/10.3390/info11060338