SICSS Festival

Tutorials, discussions, and debates about computational social science lead by the SICSS community

The SICSS Festival is a series of online events such as tutorials, panel discussions, and debates. Festival events are typically run with SICSS alumni and open everyone. The goals of the SICSS Festival are to provide new learning opportunities for people interested in computational social science, to provide a venue for community building across SICSS partner locations, and to provide an opportunity for SICSS alumni to share their expertise.

2021: Monday, June 21 - Friday, June 25

Using images and video data for social science: Challenges and opportunities

Time: Tuesday, June 22, 2021. 11am-12pm EDT

Speakers: Bryce Dietrich (Assistant Professor of Social Science Informatics at the University of Iowa), Laura Nelson (Assistant Professor of Sociology at Northeastern University), Michelle Torres (Assistant Professor of Political Science at Rice), and Han Zhang (SICSS-Hong Kong 21; Assistant Professor in Division of Social Science at The Hong Kong University of Science and Technology)

Moderator: Thomas Davidson (SICSS-Princeton 19; Assistant Professor of Sociology at Rutgers University)

Description: Social life is increasingly mediated through images and video and recent advances in machine learning make it possible to analyze these data at scale. It has become commonplace for computational social scientists to work with text and other complex, unstructured data, but we are only just seeing work that uses computational methods to study images and video. This panel focuses on this innovative area of research, exploring how scholars are using such data in cutting-edge research and the technical and ethical challenges involved in such analyses.

Open to: Unlimited registered participants. Registration has closed.

Archiving: This talk will be recorded and archived. Video


Introduction to Text Analysis in Python: A Hands-on Tutorial

Time: Tuesday, June 22, 2021. 1pm-2pm EDT

Speaker: Austin van Loon (SICSS-Princeton 19; Ph.D. student in Sociology at Stanford University)

Description: The increased availability of machine-readable text provides a unique opportunity for social scientists, granting us unprecedented access to many aspects of both historical and contemporary social life. This tutorial aims to introduce researchers to text analysis in Python, an open-source programming language. Specifically, I’ll seek to cover: using the new Twitter API v2, pre-processing text, visualizing patterns in data, and a few conceptually accessible methods for quantitatively analyzing unigram frequencies (the “dictionary method” and differential language analysis). The tutorial is divided between a publicly available Google Colab notebook (see below) which provides documented, step-by-step example code, and a live session in which I review key points from the notebook and discuss the methods explored therein more analytically.

Materials: Google Colab notebook

Open to: 45 registered participants. Registration has closed.

Archiving: This talk will be recorded and archived. Video


Panel discussion on the non-academic job market in computational social science

Time: Wednesday, June 23, 2021. 12-1pm EDT

Speakers: Dr. Aruna Balakrishnan (Head of Health & Identity Experience Research at Twitter) and Dr. Gonzalo Rivero (Associate Director of Data Labs at the Pew Research Center)

Moderator: Chris Bail (SICSS-Princeton 17, 19, 21, SICSS-Duke 18, 20; Professor of Sociology and Public Policy at Duke University)

Description: Computational social science leads to a wide range of interdisciplinary job opportunities outside academia. These panelists will share their experiences in industry settings and will offer thoughts about how an academic research program in the field could lead to a variety of fulfilling careers.

Open to: Unlimited registered participants. Registration has closed.

Archiving: This talk will not be archived.

Panel discussion on the (many) paths to computational social science research in law

Time: Wednesday, June 23, 2021. 12-1:30pm EDT, 6-7:30pm CEST

Speakers: Masha Medvedeva (Ph.D. student at University of Groningen), Jens Frankenreiter (Postdoctoral Fellow at Columbia Law School), Thales Bertaglia (Ph.D. student at Maastricht University)

Description: Three panelists represent three inspiring examples of interdisciplinary research in law. These panelists will share their experiences in learning new skills, motivations to pursue computational research, and challenges in creating innovative research agendas.

Event organized by SICSS-Law 21

Tutorial on deep learning for causal inference

Time: Thursday, June 24, 2021. 11am-12:30pm EDT

Speakers: Bernard Koch (SICSS-Los Angeles 19, 20, 21; Ph.D. student in Sociology at UCLA)

Description: This tutorial will teach participants how to build simple deep learning models for causal inference. Although this literature is still quite young, neural networks have the potential to extend causal inference to new settings where confounding is high-dimensional, non-linear, time-varying or even encoded in text, images, and networks. The tutorial will begin with 30 minutes of discussion on what deep learning is and how deep learning can be used for adjustment in observational causal inference, followed by 40 minutes of instruction on how to build your own custom deep learning models in Tensorflow 2. No prior knowledge of Tensorflow or neural networks is necessary, so this is also a great opportunity if you just want to get your feet wet with deep learning for the first time.

Expected background and computing environment: This is an advanced tutorial. Participants will need (at least two of): a basic understanding of causal inference, a basic understanding of supervised machine learning, and intermediate to advanced proficiency in some programming language to get the most out of it. No prior knowledge of deep learning is expected. The tutorial will be taught in Python, but the Tensorflow port for R is very, very similar so ideas will transfer. The tutorial will be taught in Colaboratory, an interactive notebook environment provided by Google. When you start a Colab session, Google provides you a virtual machine with a GPU, making it a great environment to get started with deep learning. Participants won’t need to install anything on their computer, but they will need a gmail or Google hosted email account to use Colab.

Materials: Tutorial; More tutorials on deep learning for causal inference

Open to: Unlimited registered participants. Registration has closed.

Archiving: This talk will be recorded and archived. Video


Taking Quantitative Description Seriously

Time: Thursday, June 24, 2021. 1-2pm EDT

Speakers: Andy Guess (Assistant Professor of Politics and Public Affairs at Princeton), Eszter Hargittai (Professor of Communication and Media Research at the University of Zurich), and Jen Pan (Assistant Professor of Communication, Political Science and Sociology at Stanford University)

Moderator: Kevin Munger (SICSS-Princeton 19; Assistant Professor of Political Science and Social Data Analytics at Penn State)

Description: We introduce the rationale for a new peer-reviewed scholarly journal, the Journal of Quantitative Description: Digital Media. The journal is intended to create a new venue for research on digital media and address several deficiencies in the current social science publishing landscape. First, descriptive research is undersupplied and undervalued. Second, research questions too often only reflect dominant theories and received wisdom. Third, journals are constrained by unnecessary boundaries defined by discipline, geography, and length. Fourth, peer review is inefficient and unnecessarily burdensome for both referees and authors. We outline the journal’s scope and structure, which is open access, fee-free and relies on a Letter of Inquiry (LOI) model. Quantitative description can appeal to social scientists of all stripes and is a crucial methodology for understanding the continuing evolution of digital media and its relationship to important questions of interest to computational social scientists.

Open to: Unlimited registered participants. Registration has closed.

Archiving: This talk will be recorded and archived. Video


Creating your own virtual lab experiment with Empirica

Time: Friday, June 25, 2021. 9am-1pm EDT

Speakers: Abdullah Almaatouq (SICSS-Princeton 17; MIT Sloan, Empirica founder & developer), Joshua Becker (SICSS-Princeton 17, SICSS-Chicago 18, SICSS-London 21; UCL School of Management, tutorial author), and Sam Dupret (UCL School of Management, tutorial author)

Description: This is a self-guided, do-it-yourself workshop that will enable you to create virtual lab experiments with Empirica. Empirica is an open-source framework that makes creating virtual lab experiments more fun and less painful. During this event, participants will first watch a set of introductory videos (about 10 mins) that provide a conceptual overview. Participants will then proceed through a written tutorial that will walk them step-by-step through building a multi-player experiment with chatrooms and automated bots. Upon completing the tutorial, participants will have created an experiment that they can further customize and that is ready to deploy to real research participants. Faculty members with extensive experience using Empirica for research—and other workshop participants—will be available in the Empirica Slack during the workshop period to help participants through the tutorial and answer more general questions, both about Empirica and related topics like subject recruitment. Outside the workshop period, the Empirica Slack remains available for general discussion and questions both with the faculty and other community members.

Expected programming background: Empirica is a platform built on ReactJS, but participants do not need to know javascript to complete the tutorial. The tutorial is designed to be self-sufficient for anyone with strong coding skills who can learn on the go. Knowledge of basic HTML is helpful.

Expected computing environment: While Empirica can run on Windows, we strongly recommend a Linux/MacOS terminal for ease-of-use and these will be made available upon request.

References:
Empirica: a virtual lab for high-throughput macro-level experiments
Empirica tutorial: Your First Experiment

Open to: Unlimited registered participants. Registration has closed.

2020: Monday, June 22 - Friday, June 26

Jump to a day: Monday. Tuesday. Wednesday. Thursday. Friday.

Monday

Panel discussion on teaching computational social science

Time: Monday, June 22, 2020. 12-1pm EDT

Speakers: Matti Nelimarkka (SICSS-Princeton 17, SICSS-Helsinki 18, SICSS-Istanbul 19, 20), Rochelle Terman (SICSS-Princeton 17), and Jae Yeon Kim (SICSS-Princeton 19, SICSS-Bay Area 20)

Moderator: Matthew Salganik (SICSS-Princeton 17, 19, SICSS-Duke 18, 20)

Description: Teaching computational social science at both the undergraduate and graduate level presents a number of challenging pedagogical questions. How should we teach a class with students from different disciplinary backgrounds? What is the role of programming in computational social science education? How should computational social science courses fit into a larger curriculum? The panelists will address these questions—and others—first in a moderated conversation and then will take questions from the audience.

Open to: Unlimited registered participants: Registration has closed.

Archiving: This talk will be recorded and archived. Video


Tuesday

Measuring cultural change in digital trace data using diversification rates

Time: Tuesday, June 23, 2020. 1-2 EDT.

Speaker: Bernard Koch (SICSS-Los Angeles 19, 20)

Description: How do we measure cultural change? Through the birth and death of “cultural lineages” (e.g., hashtags, consumer goods, or musical artists), digital trace data allows us to observe cultural change at population scale over time. This tutorial introduces a statistical framework to identify and explain culturally important historical events through changes in the diversity of cultural lineages over time. Users will learn how to run an unsupervised machine learning model called LiteRate to identify statistically significant shifts in cultural diversity, as well as more restricted models to test theoretical hypotheses about the causes of these events. The tutorial will be accessible to a broad audience.

Open to: Unlimited registered participants: Registration has closed.

Archiving: This talk will not be archived.

Discussion on diversity in computational social science

Time: Tuesday, June 23, 2020. 3:30-4:30pm EDT

Speaker: Naniette H. Coleman (SICSS-Princeton 19)

Host: Matthew Salganik (SICSS-Princeton 17, 19, SICSS-Duke 18, 20)

Description: In order to thrive, the field of computational social science needs contributions from diverse scholars. This discussion will address this broad challenge in a specific setting: the creation of SICSS-Howard/Mathematica (links to website, form to join the email list, and Facebook page), the first SICSS located at a Historically Black College or University (HBCU). Although this site has been postponed until 2021 due to COVID-19, it still serves as a useful case study for discussion. The discussion will focus on efforts to increase diversity in computational social science, why this work is important, and what should be done in the future. There will also be time for the audience to submit written questions to be asked by the host.

Open to: Unlimited registered participants: Registration has closed.

Archiving: This talk will be recorded and archived. Video


Wednesday

Computational social science to address the (post) COVID-19 reality

Time: Wednesday, June 24, 2020, 10–11:15am EDT

Speakers: Johan Bollen and Dean Eckles

Moderators: Maria Ferreira Sequeda and Monika Leszczynska (SICSS-Princeton 19, SICSS-Maastricht 20)

Description: Computational Social Science offers tools that can assist both policy makers and the private sector in the design and implementation of measures to address the economic and social challenges posed by COVID-19. How can the growing sources of (alternative) data and data science be used in a meaningful and responsible way, i.e. providing reliable and practical knowledge without compromising privacy and safety of people whose data is collected and analyzed? What are the methodological and regulatory solutions that could be applied to address the challenges and mitigate the risks? Our panelists will share their experience and ideas on these issues. There will also be enough time for questions from the audience.

Co-organized with the Maastricht Law and Tech Lab, ING, and the Think Forward Initiative

Open to: Unlimited registered participants. Registration has closed.

Archiving: This talk will be recorded and archived. Video.


Panel discussion on digital and computational demography

Time: Wednesday, June 24, 2020. 11:30am-12:30pm EDT

Speakers: Nicolò Cavalli (SICSS-Duke 18, SICSS-Oxford 19), Ridhi Kashyap (SICSS-Princeton 17, SICSS-Oxford 19), and Francesco Rampazzo (SICSS-Duke 18)

Moderator: Vissého Adjiwanou (SICSS-Princeton 17, SICSS-Cape Town 18, 19, SICSS-Montreal 20)

Description: The panel will begin with a broad introduction to digital and computational approaches to demography. Next, panelists will present examples drawing on some of their own work of digital and computational approaches to demographic research in two areas: 1) migration and 2) digital inequality and its implications for demographic processes. Following these specific examples, the panel will reflect on future directions for digital demography, both in the context of the promises and perils of this emerging area in demographic research, as well as in relation to its role within the broader computational social science community. We aim for this conversation to include a wider discussion between panelists and participants, and encourage broad participation from those within computational social science, demography and population studies communities.

Open to: Unlimited registered participants: Registration has closed.

Archiving: This talk will be recorded and archived. Video


Using Empirica for high-throughput virtual lab experiments (Session 1)

Time: Wednesday, June 24, 2020. 1-2:30pm EDT

Speakers: Abdullah Almaatouq (SICSS-Princeton 17), Joshua Becker (SICSS-Princeton 17, SICSS-Chicago 18), James Houghton (SICSS-Duke 20), and Nicolas Paton.

Description: Empirica is a new open source software platform for developing and conducting synchronous and interactive human-subject experiments. It has already been used by researchers around the world. This session will start with a live demonstration where participants will take part in a real-time experiment involving dozens of people. Next, the data from that experiment will be downloaded and analyzed. Thus, participants will have a behind-the-scenes, end-to-end experience with an experiment run with Emprica. Finally, there will be time for questions and discussion about Empirica and the future of experiments in the social sciences.

Open to: 30 registered participants. Registration has closed.

Preparatory materials: All registered participants are strongly encouraged to work through and watch Empirica videos A - E, which will require them to install software on their computers.

Archiving: See the video from the second session covering the same material.

Thursday

Creating open source software as part of an academic career

Time: Thursday, June 25, 2020. 11am-12pm EDT

Speakers: Ryan Gallagher (SICSS-Duke 18, SICSS-Boston 19), Anne Helby Petersen (SICSS-Duke 18), and Carsten Schwemmer (SICSS-Duke 18, SICSS-Bamberg 19)

Moderator: Matthew Salganik (SICSS-Princeton 17, 19, SICSS-Duke 18, 20)

Description: Almost all computational social science depends on open source software, yet very few computational social scientists actually contribute to open source software. These panelists will share their experiences developing open source software as part of an academic career, and they will offer advice for others who want to contribute to existing open source projects or start new ones. There will be time for questions from the audience.

Software by the speakers:

Ryan Gallagher

Anne Helby Petersen

Carsten Schwemmer

Open to: Unlimited registered participants: Registration has closed.

Archiving: This talk will be recorded and archived. Video


Panel discussion on the non-academic job market in computational social science

Time: Thursday, June 25, 2020. 12:30-2pm EDT

Speakers: Sudhir Venkatesh (Lead Social Scientist, Twitter & William B. Ransford Professor of Sociology & African American Studies, Columbia University), Pablo Barberá (Research Scientist, Facebook & Associate Professor, Department of Political Science & International Relations, University of Southern California), and Antje Kirchner (SICSS-Princeton 17, SICSS-RTI 19, Research Survey Methodologist, RTI International)

Moderator: Chris Bail (SICSS-Princeton 17, 19, SICSS-Duke 18, 20)

Description: Computational social science leads to a wide range of interdisciplinary job opportunities outside academia. These panelists will share their experiences in industry settings and will offer thoughts about how an academic research program in the field could lead to a variety of fulfilling careers.

Open to: Unlimited registered participants: Registration has closed.

Archiving: This event will NOT be recorded or archived.

What Can the SICSS Community Do to Recognize and Eradicate Anti-Black Racism in Computational Social Science?

Time: Thursday, June 25. 2-3 PM EDT

Organizers: Tina Law (SICSS-Duke 18, SICSS-Chicago 19) and Taylor W. Brown (SICSS-Princeton 17, SICSS-Duke 18, SICSS-Oxford 19)

Description: Right now, protests are taking place around the world in response to the murders of George Floyd, Breonna Taylor, Ahmaud Arbery, Tony McDade, Rayshard Brooks, and countless other Black lives, and many necessary conversations are being had about anti-Black racism. As members of the SICSS community, what can we do to recognize and eradicate anti-Black racism in our own field? Computational social science is a relatively new field with a lot of potential for positive social change. It is also a field born out of two long-standing disciplines–computer science and social science–that have histories of marginalizing Black scholarship and contributing to anti-Black policies and programs. What are the opportunities and challenges that we face as computational social scientists when it comes to addressing anti-Black racism? And, how can the SICSS community leverage our unique strength as a global network of teachers and scholars to combat anti-Black racism? This discussion and planning session is open to SICSS participants and alumni who are interested in actively tackling these questions together.

Open to: SICSS participants and alumni. Unlimited number of registered participants. Registration has closed.

Format: Discussion & planning session; note that this is not a panel, active participation is encouraged.

Archiving: This event will NOT be recorded or archived.

Friday

Opportunities and challenges with industry collaborations

Time: Friday, June 26, 2020. 10-11am EDT

Speakers: Dave Holtz (SICSS-Duke 18) and Sanaz Mobasseri (SICSS-Duke 18, SICSS-Boston 19)

Moderator: Matthew Salganik (SICSS-Princeton 17, 19, SICSS-Duke 18, 20)

Description: Most big data sources are controlled by companies, and many computational social scientists struggle to get access to these data. These panelists, who have experience collaborating with companies on research projects, will share insights about initiating, developing, and maintaining productive collaborations between researchers and companies. They will discuss practical issues, such as negotiating data usage agreements and navigating legal considerations for both parties, as well as describe potential risks and ethical issues created by these collaborations. There will be time for questions from the audience.

Open to: Unlimited registered participants. Registration has closed.

Archiving: This talk will be recorded and archived. Video


Using Empirica for high-throughput virtual lab experiments (Session 2)

Time: Friday, June 26, 2020. 10-11:30am EDT.

Speakers: Abdullah Almaatouq (SICSS-Princeton 17), Joshua Becker (SICSS-Princeton 17, SICSS-Chicago 18), James Houghton (SICSS-Duke 20), and Nicolas Paton.

Description: Empirica is a new open source software platform for developing and conducting synchronous and interactive human-subject experiments. It has already been used by researchers around the world. This session will start with a live demonstration where participants will take part in a real-time experiment involving dozens of people. Next, the data from that experiment will be downloaded and analyzed. Thus, participants will have a behind-the-scenes, end-to-end experience with an experiment run with Emprica. Finally, there will be time for questions and discussion about Empirica and the future of experiments in the social sciences.

Open to: 30 registered participants. Registration has closed.

Preparatory materials: All registered participants are strongly encouraged to work through and watch Empirica videos A - E, which will require them to install software on their computers.

Archiving: This talk will be recorded and archived. Video