Advances in Data Science and Environmental Archaeology

Session Type: 
Session Date and Time: 
Friday, 26 April, 2024 - 13:30 to 15:30
Primary Organizer: 
Jonathan Dombrosky - Crow Canyon Archaeological Center, Molly Carney

Data scientists combine expertise in statistical workflows, database creation and maintenance, and programming language to efficiently analyze large amounts of data. Their goal is to understand complex phenomena and offer new insights from large datasets. Modern society is now reliant on data science considering how effective it is at extracting useful information. Whether we like it or not, data science and its technologies are here to stay. Environmental archaeology—a subdiscipline with explicit ethnobiological connections—is rich in data from a vast array of different areas of research, such as geochemistry, environmental science, ecology, botany, and zoology. Environmental archaeologists have rapidly adopted data science approaches to glean accurate, precise, and replicable insights from the archaeological record. This poster session highlights new approaches in data science that help support the study of how past peoples interacted with their environments. We focus on such topics as the creation and management of ethnobiological databases, open-source tools in scientific communication, interactive data visualizations and user engagement, and new statistical and computing techniques to better understand the archaeology of human-environment interaction. Posters in this session demonstrate how these relatively new techniques and technologies are not mere distractions but offer the potential to expand our understanding of people-organism-environment interactions and relationships. Ethnobiologists can leverage data science to help sustainably support, expand, and transform the field well into the future.

Presentation format: 
Poster display (live)
, Andrew - Washington State University
, Andrea - Max Planck Institute for Geoanthropology
, Alexis - Middlebury College
, Jon - University of Oregon
, Scott - University of Oregon
, Nicole - Max Planck Institute for Geoanthropology
, Torben - Smithsonian Institution

Archaeology provides long-term records of coupled human-natural systems: the linked causes and effects of culture and environment. Archaeologists increasingly use large archaeological datasets to model human behavior in different landscapes. These data allow researchers the ability to understand human ecodynamics variability across the world and over time. We present a synthetic, global-scale archaeological database—ArchaeoGLOBE Island Extinctions database—of 4,500 events that document human and species arrivals, land use changes, and last occurrences on islands. We explore the expansion of modern humans into island environments globally to test theories about the causes of species extinctions and extirpations in these unique ecosystems using artificial intelligence methods. For example, we can use machine learning to identify which species are most likely to go extinct from human interaction. Thus, data science methods offer new insights into large archaeological and ethnobiological datasets.

Presentation format: 
Poster display (live)
, Jonathan - Crow Canyon Archaeological Center

Scholars across various fields of research have noted a reproducibility crisis. At the heart of this crisis is the fact that many results are unverifiable because researchers have not provided raw data and/or sufficient documentation. The fields of archaeology, zooarchaeology, and ethnobiology have not eluded these discussions, and they would benefit from unifying open science and publication practices. Quarto is a technical publishing system that integrates multiple programming languages and exports a range of different file types. Every figure, statistic, and calculation in a fully formatted publication or report can be tied back to raw data using one file type. Here, I present a workflow for producing technical archaeofaunal reports from active projects at Crow Canyon Archaeological Center, which are integrated with a large multisite, relational database. Quarto has enhanced internal reporting consistency, aided collaboration, reduced writing time, and increased understanding of complex archaeological phenomena.

Presentation format: 
Poster display (live)
, Molly - Oregon State University

ChatGPT and other writing tools like Grammarly and QuillBot use generative artificial intelligence (AI) language models to generate comprehensible responses and written content. All three programs work when users input questions, prompts, notes, or general queries and then ask for specific output such as paragraphs, short essays, or annotated code. These programs draw on trained data in their responses. However, all programs save input content, so human trainers and developers can review and analyze user-generated inputs before allowing content to be added to future model versions. What happens, then, when AI is “fed” cultural heritage information as an input? Who then has access to that information? How can we ethically work with archaeological or ethnobiological data and AI without compromising our commitments to descendent communities? This poster explores some of these questions and looks closely at how environmental archaeological data is archived in various AI programs and offers some initial thoughts on using AI ethically.