This week, Arianna and Anders are representing NERDS at ICWSM in Buffalo, NY, with two freshly-published papers.

  1. Narratives of Collective Action in YouTube’s Discourse on Veganism, by A. Pera and L.M. Aiello. ICWSM’24.

    We studied vegan narratives on YouTube through the lens of a theoretical framework of moral narratitves. We studied how different narratives elicit different types of responses from video commenters, and found that videos advocating social activism are the most effective at stirring reactions marked by heightened linguistic markers that relate to collective action.
  2. The Persuasive Power of Large Language Models by A.G. Møller and L.M. Aiello. ICWSM’24.

    Can artificial agents interact with each other to reproduce human-like persuasive dialogue? And do the arguments they generate sound persuasive to humans? We used Llama2 to test different persuasion strategies, and asked humans to rate them. We found that arguments that included factual knowledge, markers of trust, expressions of support, and conveyed status were deemed most effective according to both humans and agents.

NERDS at Como Summer School and WebSci’24

Arianna and Anders participated to the first editions of the Computational Social Science Summer School in Como, presenting their work on the COCOONS project. Arianna, Daniele, and external collaborator Maddalena Torricelli also attended the WebSci conference in Stuttgard, presenting an analysis of climate action communication on TikTok [paper], the use of hypergraphs to model opinion dynamics in large-scale social media [poster], and the role of interfaces in shaping human creativity during the interaction with generative AI tools [paper].

New NERDS paper out on Machine Learning in Humanitarian Work

We published a new paper:

THE OPPORTUNITIES, LIMITATIONS, AND CHALLENGES IN USING MACHINE LEARNING TECHNOLOGIES FOR HUMANITARIAN WORK AND DEVELOPMENT, by V. Sekara, M. Karsai, E. Moro, D. Kim, E. Delamonica, M. Cebrian, M. Luengo-Oroz, R. Moreno Jimenez, M. Garcia-Herranz, published in Advances in Complex Systems

Novel digital data sources and tools like machine learning (ML) and artificial intelligence (AI) have the potential to revolutionize data about development and can contribute to monitoring and mitigating humanitarian problems. The potential of applying novel technologies to solving some of humanity’s most pressing issues has garnered interest outside the traditional disciplines studying and working on international development. Today, scientific communities in fields like Computational Social Science, Network Science, Complex Systems, Human Computer Interaction, Machine Learning, and the broader AI field are increasingly starting to pay attention to these pressing issues. However, are sophisticated data driven tools ready to be used for solving real-world problems with imperfect data and of staggering complexity? We outline the current state-of-the-art and identify barriers, which need to be surmounted in order for data-driven technologies to become useful in humanitarian and development contexts. We argue that, without organized and purposeful efforts, these new technologies risk at best falling short of promised goals, at worst they can increase inequality, amplify discrimination, and infringe upon human rights.

Postdoc wanted for analyzing networks of the Roman Empire

We are looking for a postdoc to work on a cool multidisciplinary project on networks and the Roman Empire ⚔️, at the NERDS (NEtwoRks, Data, and Society) research group at IT University of Copenhagen. Apply here if interested (deadline May 23rd)! 

In the project, you will be expected to develop network analysis techniques to work with extremely incomplete and heavily biased archaeological data. The idea is to try to reconstruct social relationships between different places in the Roman Empire. We are going to integrate remains at various sites with the most detailed to date reconstruction of Roman mobility. Check out this cool Roman roads network picture. That’s the stuff you’re going to work with!

The work is part of a Villum Synergy project. You’ll be supervised directly by Michele Coscia. You will interact on a regular basis with a team of cool archaeologists from Aarhus University, headed by Tom Brughmans.

Here’s the link to read more about the call and apply (deadline May 23rd):

The NERDS group is a down-to-earth and fun place to be. Copenhagen is often named as the best city in the world to live in, and for good reasons. It’s world-renowned for food, beer, art, music, architecture, the Scandinavian “hygge”, and much more. In Denmark, parental leave is generous, and child-care is excellent and cheap.

New NERDS paper out on success in tennis

We published a new multi-NERDS paper, concluding a successful previous internship of Chiara Zappalà!

Early career wins and tournament prestige characterize tennis players’ trajectories, by C. Zappalà, S. Sousa, T. Cunha, A. Pluchino, A. Rapisarda and R. Sinatra, published in EPJ Data Science

We study the unfolding of tennis players’ careers to understand the role of early career stages and the impact of specific tournaments on players’ trajectories. We employ a comprehensive approach combining network science and analysis of the Association of Tennis Professionals (ATP) tournament data and introduce a novel method to quantify tournament prestige based on the eigenvector centrality of the co-attendance network of tournaments. Focusing on the interplay between participation in central tournaments and players’ performance, we find that the level of the tournament where players achieve their first win is associated with becoming a top player. This work sheds light on the critical role of the initial stages in the progression of players’ careers, offering valuable insights into the dynamics of success in tennis.

5 years of NERDS!

NERDS was founded 5 years ago by three young assistant professors and one associate professor, to be a reference point at ITU for the research on network and data science applications to social systems. At 5 years old, we have learned to write our name (we have a logo), to follow rules, and to use a fork and knife for eating. And hoo boy, did we use that fork last friday when celebrating our anniversary with an original NERDS cake!

We also got new group photos taken (by Sebastian Mateos Nicolajsen – thx!), see below, because we have grown to 20+ members over the years! For all bean counting aficionados, we also won over 5M EUR of research funding and published 84 papers so far.

Further highlights, shown in the timeline above. We:

Looking back to our goals 5 years ago, we have all reason to be proud to have 1) built up a flourishing network of Denmark-based network/data science research groups, connecting ITU, KU, DTU, and others, 2) successfully impressed several funding agencies and public stakeholders to engage with us solving societal problems with our research. We will continue along this road, developing further our group in a safe and fun environment.

In our near future, we look forward to welcoming several new group members in the fall, including one assistant professor and several PhDs/Postdocs.

Live long and prosper 🖖
Luca, Luigi, Roberta, Claudia, Anastassia, Jacob, Vedran, Sandro, Anders, Michele, Anders 2, Ane, Toine, Mesut, Luca 2, Michael, Arianna, Clement, Elisabetta, Alessia, Jacopo, Daniele, Nicoló

Mesut Kaya has joined NERDS

We are happy to welcome Mesut Kaya to our research group!

Mesut joins us as an Industrial Postdoc funded by the Innovation Fund. Previously, Mesut was at Aalborg University Copenhagen; before in Delft University and University College Cork. Mesut works on recommender systems in general, at NERDS he will be working with Toine Bogers specifically on recommendation in the HR domain: job recommendation, candidate recommendation, and now the fairness of algorithmic hiring in general.

New NERDS paper out on bicycle network quality in Denmark

We published a new all-NERDS paper, applying our BikeDNA tool to the whole country of Denmark as part of our Cykelpulje project!

How Good Is Open Bicycle Network Data? A Countrywide Case Study of Denmark, by A. Rahbek Vierø, A. Vybornova, and M. Szell, published in Geographical Analysis

We compare the two largest open data sets on dedicated bicycle infrastructure in Denmark, OpenStreetMap (OSM) and GeoDanmark, in a countrywide data quality assessment, asking whether the data are good enough for network-based analysis of cycling conditions. We find that neither of the data sets is of sufficient quality, and that data conflation is necessary to obtain a more complete data set. Our analysis of the spatial variation of data quality suggests that rural areas are more prone to incomplete data. We demonstrate that the prevalent method of using infrastructure density as a proxy for data completeness is not suitable for bicycle infrastructure data, and that matching of corresponding features is thus necessary to assess data completeness. Based on our data quality assessment, we recommend strategic mapping efforts toward data completeness, consistent standards to support comparability between different data sources, and increased focus on data topology to ensure high-quality bicycle network data.

Explore also the interactive map:

GrowBike.Net covered in Austrian media

Our project GrowBike.Net was just covered in Austrian media, both in an online news report and on Austrian TV station ORF 2, in the news report Wien heute. If you understand German, watch our 15 seconds of fame here:

The written report Radwege sollen möglichst direkt verlaufen:

GrowBike.Net is an interactive platform resulting from an ITU master project, visualizing the results of our 2022 paper Growing urban bicycle networks: The idea is to simulate the creation of a cohesive bicycle network inspired by the Dutch CROW design manual for bicycle traffic. Studying these synthetic networks informs us about the geometric limitations of urban bicycle network growth and can lead to better designed bicycle infrastructure in cities. In the case of the Austrian news report, the key metric of directness was highlighted: Bicycle networks should allow for direct paths without substantial detours.

Five new NERDS publications out!

We have been very productive this year already! Five new NERDS publications are released this week:

  1. Which sport is becoming more predictable? A cross-discipline analysis of predictability in team sports, by M. Coscia, published in EPJ Data Science

    We analyze more than 300,000 professional sports matches in the 1996-2023 period from nine disciplines, to identify which disciplines are getting more/less predictable over time. We investigate the home advantage effect, since it can affect outcome predictability and it has been impacted by the COVID-19 pandemic. Going beyond previous work, we estimate which sport management model – between the egalitarian one popular in North America and the rich-get-richer used in Europe – leads to more uncertain outcomes. Our results show that there is no generalized trend in predictability across sport disciplines, that home advantage has been decreasing independently from the pandemic, and that sports managed with the egalitarian North American approach tend to be less predictable. We base our result on a predictive model that ranks team by analyzing the directed network of who-beats-whom, where the most central teams in the network are expected to be the best performing ones.

  2. Algorithmic Fairness: Learnings From a Case That Used AI For Decision Support, by V. Sekara, T.S. Skadegard Thorsen, and R. Sinatra, published by the Crown Princess Mary Center

    This policy brief provides a small introduction to algorithmic fairness and an example of auditing fairness in an algorithm which was aimed at identifying and assessing children at risk from abuse.

  3. The Parrot Dilemma: Human-Labeled vs. LLM-augmented Data in Classification Tasks, by A.G. Møller, J.A. Dalsgaard, A. Pera, L.M. Aiello (accepted at EACL’24).
    How good are Large Language Models in generating synthetic examples for training classifiers? To find out, we used GPT4 and Llama2 to augment existing training sets for typical Computational Social Science tasks. Our experiments show that the time to replace human-generated training data with LLMs has yet to come: human-generated text and labels provide more valuable information during training for most tasks. However, artificial data augmentation can add value when encountering extremely rare classes in multi-class scenarios, as finding new examples in real-world data can be challenging. 

  4. Shifting Climates: Climate Change Communication from YouTube to TikTok, by A. Pera, L.M. Aiello (accepted at WebSci’24).

    How do video content creators tailor their communication strategies in the era of short-form content? We conducted a comparative study of the YouTube and TikTok video productions of 21 prominent climate communicators active on both platforms. We found that when using TikTok, creators use a more emotionally resonant, self-referential, and action-oriented language compared to YouTube. Also, the response of the public aligns more closely to the tone of the videos in TikTok.

  5. The role of interface design on prompt-mediated creativity in Generative AI, by M. Torricelli, M. Martino, A. Baronchelli, L.M. Aiello (accepted at WebSci’24).
    We analyze 145k+ user prompts from two Generative AI platforms for image generation to see how people explore new concepts over time, and how their exploration might be influenced by different design choices in human-computer interfaces to Generative AI. We find that creativity in prompts declines when the interface provides generation shortcuts that deviate the user attention from prompting.