Materials for the DH 2019 Workshop


Welcome to the Time Series workshop being offered at DH2019 by the Culture Analytics group! Much of the materials you will need for the workshop are available here, as will be notes, handouts, and possibly slides from the various talks. If you are viewing the GitHub pages version of the repository, you will want to browse to the repository itself to see everything as well as to make it easier to download or clone the contents of the repository.

If you enrolled in the workshop, please be sure to fill out the survey available on Google Drive (see email) so that we can prepare the best possible experience for everyone. Our goal is to expand the number of researchers who consider themselves to be working within the arena of culture analytics. The survey simply asks you to tell us about you: name, affiliation, research interest(s), and, if you have a link to a paper or a project description you would like others to know about.

If you are curious about some of the history of the culture analytics group, an included whitepaper offers some insight.


If you are in the workshop, then you know it is due to last much of the day, Tuesday, July 9 from 9:00 to 16:00. In order to make the day not feel long, however, we have scheduled a number of breaks as well as kept the talking to a minimum: we have tried to make the workshop as hands-on and focused on your work as possible.

Please note that the schedule was updated on June 30 to reflect changes in presenter availability as well as to sync the timings for breaks and lunch with the published conference schedule.

Time Event
9:00 Welcome and Introductiopn. Followed by introduction of participants. Clarisse Bardiot and Mila Oiva.
9:30 What is Culture Analytics? Our keynote talk is designed to give a quick definition of the domain, with a view to the multi-stream history of research in this area, to establish the nature of time series, and to highlight recent research by various individuals, focused on time series in culture analytics. John Laudun.
10:30 Morning Break.
11:00 Collection-Scale Visualization with PixPlot PixPlot is open-source software that presents tens of thousands of images organized according to visual similarity. We’ll talk about prior work in this area, explore how computers ‘see’, experience some existing collections in PixPlot, and future directions for the tool. Peter Leonard
11:30 Automated Movement and Choreography Analysis of Video Data via Deep Learning Pose Detection. This tutorial will explore the possibilities for pose and movement analysis available via deep learning-based tools for pose extraction from video. Participants will have the opportunity to experiment with computational approaches to characterizing and comparing extracted poses using R Studio and online Python notebooks. Code and data [SLIDES] Peter Broadwell.
13:00 Lunch
14:00 Visualizing Time Series with R Using classical methods in time series analysis and exploratory data analysis, we show how various temporal datasets can be efficently analyzed and dispalyed. The focus will be on understanding how these methods differ and seeing various applications to cultural analytics. Participants with prior programming experience are encouraged to follow-along with the analysis using the open source programming language R. Code and data Taylor Arnold.
14:45 Afternoon Break
15:15 Visualizing Time Series (continued).
15:45 Workshop Wrap-up and Discussion


Arnold: University of Richmond • Bardiot: Université de Valenciennes • Broadwell: Stanford University • Laudun: University of Louisiana • Leonard: Yale University • Oiva: Turun Yliopisto (University of Turku).