On March 10th, the CSSLab celebrated one year of working to define the new field of computational social science. To mark its anniversary month, the Lab’s Executive Director, Valery Yakubovich, reflects upon our mission to adapt the core principles of CSS research to the Lab’s daily operations.
The CSSLab’s first year is behind us. By all measures, it was a good formative year. Our researchers published sixteen articles, including six in major interdisciplinary journals: four in the Proceedings of the National Academy of Sciences, one in Nature, and one in Management Sciences. A generous gift from Richard Mack allowed us to launch the PennMAP initiative. The Lab now includes six full-time researchers, seven PhD students, and a rapidly growing network of collaborators across the US and Europe. Over the course of the year, we employed more than thirty undergraduate and graduate research assistants. A team of six staff members support these activities, as well as the global effort to reshape the social sciences through the newly created International Society for Computational Social Science and its flagship conference.
As we’ve made these accomplishments, one persistent question for me has been: how does one organize work in a new lab that aspires to create a new field? I believe that the answer lies in aligning the lab with the field in both theory and practice. Since we at the CSSLab study socio-technical systems, from work teams to information ecosystems, we should see the Lab as a socio-technical system in its own right, organized according to the same basic principles of computational social science by which our researchers organize their projects. Slightly modified for the Lab’s internal purposes, these principles are: use-inspired actions, infrastructure, collaboration, and partnerships.
Use-inspired actions are thoughtful responses to the Lab’s “pain points”—real problems we have to address, not something only staff are interested in doing for one or another idiosyncratic reason. Since we are a fledgling lab in a fledgling field, such actions are often exploratory and improvisational. However, if they become repetitive, we can invent standard business processes and tools and automate them as much as possible to ensure reliability and replicability. So far, our priority has been to standardize internal task assignment and coordination, RA hiring and onboarding processes, and protocols and templates for data providers, researchers from other institutions, and industry partners.
Computational social science is impossible without a digital infrastructure including platform and data components. With this in mind, we strive to adapt computational methods to our everyday operations. We are a paperless Lab which stores its data on three main digital platforms: AWS, GitHub, and Google Drive. Our team cross-references these platforms and manages our organizational tasks and projects on GitHub the same way our researchers manage theirs during the research process. Accordingly, standardized business processes turn into computer code and algorithms implemented using a variety of tools such as GitHub issues and Excel macros.
Computational social scientists engage in collaboration to overcome interdisciplinary silos and develop shared concepts and generalizable theories. To make such collaborations easier, the Lab launched a research network that secures access to our infrastructure to scholars from around the world who run projects with our researchers, contribute to our infrastructure, or bring funds to run their own projects.
Internally, collaboration between researchers and staff on specific research projects is the best way to erase our internal boundaries by creating a common language, understanding, and experiences. Since our main output is research papers, the staff see themselves primarily in a supporting role vis-a-vis researchers. This said, there are projects where the staff become equal partners, for example, in designing and implementing data pipelines. And the principle of use-inspired research requires the conversion of our research output into practical solutions to specific societal problems—a key area where staff should be capable of taking a leading role.
Academia-industry partnerships are an area where the Lab’s staff are already playing such a role. To streamline our research process, we took over the Lab’s relationships with data providers, digital platforms, and industrial research sites. Some of these relationships contribute to and support our own research infrastructure; others open access to the infrastructure of our partners. Since computing and data costs are often prohibitive, such partnerships are critical for the Lab’s success. The question is: Why would industry partners be interested? The answer brings me back to the first principle: use-inspired research means that the Lab contributes to basic science not for the sake of science per se, but for the sake of addressing important problems, including those important to our partners. If, in the midst of everyday preoccupations, we pause and ask ourselves what exact problem we are trying to solve, better actions, decisions, and processes will follow.
And there are plenty of problems to focus on: designing novel pipelines for PennMAP, running high-throughput experiments, conducting field studies with industry partners, winning competitive research grants. I wish that, while working towards these goals, our Lab will grow into something much more than the sum of its parts, where our team efforts add value above and beyond our individual accomplishments. For me, this is the ultimate goal for our second year.