IC2S2 2023 Copenhagen
International Conference on Computational Social Science
IC2S2 launched with flying colours in 2015 in Helsinki, where I was a PhD student then and a volunteer. The first event attracted the top scientist in social science and social computing.
Fast-forward to 2023 in Copenhagen, and it was another blast event, proving that IC2S2 has grown to be the dominant and deservedly named premier conference at the intersection of social and computational science.
The Nokia Bell Labs, Cambridge team I am part of (Social Dynamics), had a strong presence at this year’s conference with: 1 tutorial (Computation Social Science from Space), 3 talks (Positive Stress in Companies, The Online Medical Taxonomy, and Topics of Our Dreams), and 3 posters (Multidimensional Tie Strength and Economic Development, Understanding health care demands through social media engagement, and Combining Survey Items and Digital Traces to Measure Workplace Depression) presented by our members.
The Keynotes included Sharad Goel of Harvard, who talked about why we should not include all the available factors to maximize the predictive performance of our models, at least not when the target of prediction differs from the true outcome of interest (e.g., when predicting healthcare costs as a proxy for healthcare needs). He then went on to the issue of using race to predict diabetes prevalence, which is one of the most commonly employed control variables for it. By first showing us that, indeed, racial minorities, especially Asian, are substantially more likely to have diabetes, and then still providing a breakdown analysis revealing that the benefits of actually using this protected attribute are minor, he took us through an entire roller coaster of ideas. He even asked us for opinions, and we were mostly wrong.
Another keynote was by Stefan Gössling, who discussed the period of travel maximization, social drivers of transport demand, and travel influencers in the context of climate change. Some in the audience asked, possibly feeling partly guilty for having travelled there as I did, but how should we do it if attending and presenting at conferences is one way for us to progress in scientific careers? Prof. Gössling gave a neutral and kind response to this one of the many challenging questions he tackled.
Among the many inspiring plenary talks, I will start from the one that later was awarded as the best: Quantifying attention via dwell time and engagement in a social media browsing environment by Ziv Epstein, Hause Lin. They developed an ecologically valid social media research tool and showed that the content on which people dwell is substantively different from the content that they actually share. For me, and possibly others, it was impressive that we all have studied social media and social networks for so long already, and nobody has tested this ‘simple’ looking idea.
Another talk that made an enormous impression and also saddened me was from Tara Sowrirajan, who showed that by just being women, we are likely to be penalised for boundary-pushing innovation, while men are at the same time rewarded for it. And women, on average, are also delayed more and denied more patents.
Continuing on the topic of females and work, there was an excellent talk by a high-school student Zeyneb Nahide Kaya, who used the Enron dataset and found differences in the email communication directed to women compared to males at work.
I look forward to another inspiring and fruitful event in Philadelphia next year.