Culture Analytics Workshop : time series

This one-day workshop offers an exploration of culture analytics, aimed at an audience of students and scholars interested in understanding the intersection of analytics and the humanities. It follows a three-year program organized by Institute for Pure and Applied Mathematics (IPAM) at UCLA, gathering a numerous and lively community of people coming from humanities, media studies, computer science and mathematics.

Culture Analytics is, by definition, a collaborative, translational data science that explores culture and cultural interaction as a multi-scale / multi-resolution phenomenon. The macroscopic view, that allows a researcher to move from the microscale of close reading, up through the mesoscales, and on to the macroscale of distant reading, is a hallmark of the discipline. Culture analytics as a field is focused on the productive intersection between humanities, mathematics, and data science.

Researchers from the Humanities, the Social Sciences, the Mathematical Sciences, and the Data Sciences are now collaborating to identify, document, and integrate concepts, methods and tools that will provide an intellectually and ethically sound approach to the study of cultures across time and across space, leveraging the enormous gains made in the past decade in computation and machine readable cultural archives, from libraries and museum collections to the born digital cultural expressions of billions of people on the internet. This rapid proliferation of digital data has made the role of Culture Analytics all the more central particularly given the potential for significant benefit that lies in harnessing the domain expertise of researchers across these disciplines.

Time series are a very important issue that can be addressed with culture analytics methodologies. People working with digitized archives, newspapers, huge dataset of images are interested to discover evolutions, patterns, trends over time. This kind of research questions requires specific computational approaches. This workshop after a general introduction about culture analytics (Keynote by John Laudun, one of principal participants of the culture analytics long program at IPAM/UCLA) will present different study cases, followed by other short presentations. After the introductory part the program will be consecrated to a tutorial on time series in scale.

For this workshop, the target audience within the digital humanities community are those scholars interested in utilizing more mathematical and algorithmic approaches in the analysis of cultural data. The expected number of participants is 25 persons.

For the afternoon part of the workshop the participants will be sent instructions how to download the needed program and dataset to their laptops. The participants will bring their own laptops to the workshop.

Event Flow: Moments of Innovation, Disruption, and Reflection in Public Discourse - Melvin Wevers (Utrecht University) & Kristoffer Nielbo (Aarhus University)

Tracing collaboration over time in the Leonardo journal - Clarisse Bardiot (Université Polytechnique Hauts-de-France), Peter Broadwell (Stanford), Maria d’Orsogna (California State University at Northridge), Mila Oiva (University of Turku), Pablo Suarez (UNAM), Timothy Tangherlini (UCLA), Melvin Wevers (Utrecht University)

Measuring and Presenting time series of Astrophotography in Science, Public Discourse and (Virtual) Museums Exhibitions - Ekaterina Lapina-Kratasyuk (National Research University Higher School of Economics & Leon Gurevitch (Victoria University of Wellington)

Spreading News Globally in the 19th century press - the Oceanic Exchanges - & Mila Oiva (University of Turku)

Large images dataset overtime : PixPlot new features - Peter Leonard (Yale University)

Audio and video time series analysis - Peter Broadwell (Stanford University) & Timothy Tangherlini (UCLA)

This hand-on tutorial offers participants a chance to explore how time series analysis can be used both to examine a single text and to examine a corpus. Examples include a short story, a novel, and a corpus of newspapers. Exercises include parsing texts in various ways and then deriving values through topics and sentiment and then understanding change over time through the Hurst exponent. Workshop participants will need to be familiar with Python or R, or at least interested in becoming so. Interactive possibilities, including application of data from participants, will be made possible through a Jupyter notebook—the default installation of Anaconda is acceptable.

Organizing Committee members

The organizing committee members were part of the core group of the Culture Analytics Long Program at IPAM (UCLA).

John Laudun is Doris H. Meriwether/BORSF Endowed Professor of English at the University of Louisiana. His research focuses on folk narrative, both as a textual production in and of itself as well as a networked phenomenon. His published work includes a book-length study of embedded creativity and articles on folklore in both traditional and digital environments. In addition to scholarly journals and anthologies, his work is also featured in archives, CDs, films, and television series, and he has received funding from the NSF, NEH, the Andrew Mellon Foundation, the MacArthur Foundation, and the U.S. Department of Education. His work on textual analytics with Python has been featured at CERN, Duke University, among others.

Email: jlaudun@me.com

Mila Oiva, postdoctoral researcher, University of Turku, Finland. Mila Oiva (PhD) is Cultural Historian and expert on Russian and Polish 20th century history and digital humanities. She works as a postdoctoral scholar at the Trans-Atlantic Platform/Digging into Data funded Oceanic Exchanges project at the University of Turku. She is currently co-editing a special issue “Lab & Slack. Situated Research Practices in Digital Humanities” to Digital Humanities Quarterly and an edited volume “History in the Digital Era” for the Helsinki University Press. Her research interests consist of transfer of knowledge and information flows and temporal change. Email: milaoiv@utu.fi

Ekaterina Lapina-Kratasyuk, PhD, associate professor, National Research University Higher School of Economics (Moscow, Russia). Ekaterina Lapina-Kratasyuk is a professor and researcher in Media & Cultural Studies with an especial focus on Culture Analytics in field of spatial studies (urban studies, astroculture analyses). She is the co-editor (with Evgenia Nim and Oxana Moroz) of Tuning Language: Communication Management in Post-Soviet Space (2016); co-editor (with Oxana Zaporozhets) of Interactive City: Urban Life in New Media Age (2019), and author of numerous papers on digital humanities, new media, spatial media and transmedia. She is also a head of a MediaSpace research and educational program in the Space Museum in Moscow, Russia. Email : kratio@mail.ru

Clarisse Bardiot, PhD, associate professor, Université Polytechnique Hauts-de-France (France) and research fellow at the CNRS. Clarisse Bardiot is a researcher, a publisher and a curator, who has been working in the fields of digital performances history, digital humanities, documentation and preservation of time based media art works, and digital publishing. Since 2012, she has received various grants to develop Rekall and MemoRekall, an open source environment and a webapp to analyse, document and preserve time based media art.

Workshop Facilitators

Kristofer Nielbo is associate professor of humanities computing at University of Southern Denmark, where he runs an eScience unit for the humanities in the SDU eScience Center. KLN has specialized in applications of quantitative methods and computational tools in analysis, interpretation and storage of cultural data. He has participated in a range of collaborative and interdisciplinary research projects involving researchers from the humanities, social sciences, health science, and natural sciences. His research covers two areas of interest of which one is more recent (automated text analysis) and the other (modeling of cultural behavior) has followed him during his entire academic career. Both interests explore the cultural information space in new and innovative ways by combining cultural data and humanities theories with statistics, computer algorithms, and visualization.

Taylor Arnold is assistant professor of statistics at the University of Richmond where he is the co-director of the Distant Viewing Lab. He studies massive cultural datasets in order to address new and existing research questions in the humanities and social sciences. He specializes in the application of statistical computing to large text and image corpora. The study of data containing linked text and images, such as newspapers with embedded figures or television shows with associated closed captions, is of particular interest. Dr. Arnold's work has been funded by the National Endowment for the Humanities, American Council of Learned Societies, Defense Advanced Research Projects Agency, and the Collegium de Lyon's Institut d’études avancées. He has authored the books Humanities Data in R (Springer, 2015) and A Computational Approach to Statistical Learning (CRC Press, 2019).

Ross Deans Kristensen-McLachlan is a Research Assistant based at Aarhus University. His academic background is in English Language and Linguistics but he now works more broadly in computational humanities and cultural analytics. He has a strong interest in English historical linguistics, as well as cognitive and computational approaches to lexical semantics and textual analysis. Most recently, he has been involved in the Digital Literacy project at Aarhus, where he provides digital support for a number of research projects working with a diverse range of texts – from 17th-century English drama and Stephen King novels, through to Facebook groups and contemporary Danish church sermons.