A collection of post-mortems
Table of Contents
Published on: December 3, 2025
We’ve just completed the 2025 Summer Institutes in Computational Social Science. The purpose of the Summer Institutes is to bring together graduate students, postdoctoral researchers, and beginning faculty interested in computational social science. The Summer Institutes are for both social scientists (broadly conceived) and data scientists (broadly conceived). This summer we had a mixture of in-person institutes and virtual institutes across the world. In addition to various partner locations run by SICSS alumni, we continue to grow our network of computational social science researchers.
These post-mortems describe, for each site, a) how the Institute was run, b) what each site thinks worked well, and c) what each site will do differently next time. We hope that this will be useful to others organizing similar Summer Institutes or future organizers of SICSS sites. If you are interested in hosting a partner location of SICSS 2026 at your university, company, NGO, or governmental organization, please read our information for potential partner locations.
This page includes post-mortem reports from all locations in order to facilitate comparisons, as well as an overview of key themes and takeaways. As you will see, different sites did things differently, and we think that this kind of customization was an important part of how we were successful.
Skip to:
Our organizing team consisted of five members: the head of the host department, his administrative assistant, and three PhD students. The three students were responsible for communicating with guest lecturers, selecting participants, and organizing the day-to-day activities during the summer school.
The outreach and application process was overall smooth and effective. We received 35 applications for 20 places. The three student organizers independently reviewed the CVs and cover letters, scored each candidate based on the relevance of their research interests and skills, and discussed their comments on each application. Out of 20 selected participants, four cancelled their applications, and 16 participants completed the program. To help applicants make informed decisions, we published a tentative schedule early, allowing prospective participants to gain a clear idea of the program structure before applying. We also included preparatory reading materials for some of the lectures, when suggested by the lectures. However, participants later indicated that providing such materials for all sessions would have been beneficial.
The first week was dedicated to lectures, tutorials, and keynote sessions delivered by leading experts in computational social science and related fields. Highlights included Savvas Zannettou’s lecture on multimodal analysis with CLIP, Emilio Zagheni’s session on biases in digital trace data, Abhisek Dash’s case study on content moderation in decentralized platforms, and Bernie Hogan’s introduction to Network Canvas and network analysis methods. Participants also gained practical experience through hands-on activities with tools such as LinkedIn Campaign Manager, Gephi, and NetworkX. Complementing the technical training, keynote speakers such as Uwe Albrecht and Georg Wenzelburger addressed pressing societal issues, from extremism to the democratic implications of AI. To close the week, participants engaged in a lively research speed dating activity, which allowed them to connect around shared research interests and form groups for the upcoming collaborative projects in the second week. Regarding the schedule, participants mentioned that the first week felt somewhat overwhelming, as there were lectures: hands-on activities, and keynotes almost every day. For future editions, we plan to reduce the number of keynotes and shift some to week two to ensure a more balanced load.
On Saturday, we promoted a social event with a cultural visit to the Völklinger Hütte, a UNESCO World Heritage site, followed by a communal dinner that reinforced the collaborative spirit of the event. Having multiple social events throughout the two weeks allowed participants to interact with one another and the organizers, which contributed to a friendly and easy-going atmosphere.
The second week transitioned from structured training to independent group projects, where participants applied the knowledge and methods acquired during the first week to their own research questions. The week began with lectures by Daniela Braun, Rosa Navarrete, Giuseppe Carteny, and Alexander Hartland on computational approaches to political science, as well as a keynote by Ridhi Kashyap on gender inequalities in the digital world. Participants then worked intensively in small teams (2-4 people), developing and refining projects that combined innovative data collection and analysis strategies with substantive social science questions. We encouraged participants to brainstorm and share their own ideas before presenting the ideas proposed by the lecturers. This encouraged active engagement, and all groups ultimately worked on their own projects and data collection rather than relying on the pre-provided ideas and datasets. Additional keynotes by Kevin Baum on ethical reasoning in AI and Ingmar Weber on the use of satellite imagery in social science research provided further inspiration. On the final day, five project teams presented their results, showcasing the creativity, technical expertise, and interdisciplinary collaboration that had been fostered throughout the program. The event concluded with a farewell gathering at the Ausländer Café on the Saarland University Campus.
To evaluate the success of SICSS Saarbrücken 2025 and gather input for future editions, we distributed a feedback form to all participants at the conclusion of the program. The form combined quantitative questions (e.g., overall satisfaction, usefulness of lectures and tutorials, logistics, and organization) with open-ended questions (e.g., most valuable aspects, areas for improvement, and suggestions for future topics). The results showed very high levels of satisfaction with both the academic content and the organization of the summer school. Participants especially valued the balance between lectures, hands-on tutorials, and group project work, as well as the diversity of speakers and their interdisciplinary perspectives. The participants also highlighted areas for improvement, including a desire for more involvement from professors during the group project phase and final presentations.
The organizing team met to review the feedback received in detail and agreed to incorporate these insights into the planning of future editions. Based on both participants’ feedback and logistical considerations, the decision was also made to host SICSS Saarbrücken as a biennial event, ensuring high-quality programming while reducing the risk of low quorum.
We’ve divided the post-mortem into 5 parts: 1) outreach and application process, 2) pre-arrival and onboarding, 3) the first week, 4) the second week (group projects) and 5) post-departure.
The third edition of SICSS in Poland, organized as SICSS-AMU/Law, drew extensively on the experience of the previous 2022 and 2024 editions. While those earlier editions were held in an online format, the 2025 edition took place entirely in person, which introduced both new opportunities and new organizational challenges. Six co-organizers of the previous editions were again involved in the preparation of the school, and we recruited four new people (2 responsible for outreach and marketing, and 2 Teaching Assistants). Already at the recruitment stage, we were guided by the lessons learned from the previous editions and composed a team to improve those areas, where we saw room for improvement.
The first change concerned the school’s advertising strategy. The marketing approach utilized in the previous edition proved highly effective: we worked with people we already knew, trusted, and could rely on. Because this approach worked so well, we decided to keep the same model for 2025 and simply expand it by adding a few additional trusted team members. This further improved the quality of our promotional work and allowed us to develop a much more diverse marketing output. Throughout the pre-institute period, we produced a wide range of content, including fun facts, highlights of interesting events, and posts showcasing the skills and knowledge participants could gain during SICSS-AMU/Law.
As in the previous edition, our plan was to gather young scientists from Poland and surrounding countries through an extensive advertising campaign, for which we used not only social media, but also e-mail. Following the experiences of the previous editions, we decided to continue prioritizing LinkedIn as our main communication channel. In 2025, we increased both the frequency and variety of posts, regularly sharing updates about invited speakers, the evolving schedule, recommended readings, and introductions of our team members. We also expanded our content strategy by adding short educational posts. This allowed us not only to keep potential applicants informed but also to build stronger engagement and visibility within the broader social science and legal tech communities. Another change we introduced was content planning – we managed to make a list of possible ideas for social media posts and schedule the posts so we did not have to brain-storm ideas every single time there needs to be something posted online. One thing we learned is that the more ambitious we want our social media to be, the more creative work there is to be done. We wanted to keep everything integral as far as esthetics (templates, logos) are concerned. This generated much more work, but it seems it paid off well.
The traditional promotion channels — the University website, LinkedIn, and mailing lists — proved highly effective in the 2025 edition. Thanks to the stronger and more coordinated marketing strategy, as well as the improved quality and consistency of our content, we were able to reach a sufficient number of applicants without any difficulties. As a result, unlike in previous years, we did not need to rely on paid advertising and ultimately decided to abandon it altogether.Our organic outreach was strong enough to attract a diverse and well-qualified pool of candidates. During the period of our social-media activity for the 2025 edition, the number of followers of our communication channels increased by 92.7%, demonstrating both growing visibility and sustained interest in the SICSS-AMU/Law initiative. We also conducted an intensive mailing campaign, as it is the only direct channel of communication with potential candidates. For that purpose, we used a database of e-mail addresses from selected countries based on publicly available sites, which was created as a part of the previous edition. In the effect, for three weeks we have sent around 10.000 e-mails with invitations to the Institute.
As a result, we received around 65 applications ranging from the US to Europe. That was enough to run an institute and make some selections. In order to select participants we used the following criteria: 1) research and teaching in the area of computational social science 2) contributions to public goods, such as creating open-source software, curating public datasets, and creating educational opportunities for others 3) likelihood to benefit from participation 4) likelihood to contribute to the educational experience of other participants, and 5) potential to spread computational social science to new intellectual communities and areas of research. Further, when making our evaluations, we accounted for an applicant’s career stage and previous educational opportunities. A factor that was eventually hard to assess is digital skills.
Finally, we selected 25 participants representing universities from Latvia, Germany, France, Slovenia, Greece, Switzerland, and Poland, as well as various public administration offices. Finally, 16 participants arrived in person and all of them completed this year’s SICSS-AMU/Law. This outcome confirmed once again that it is advisable to accept a slightly larger group at the outset, anticipating that some participants may ultimately not be able to attend as they can struggle with securing enough financing for in-person participation.
Alongside the promotion and participants’ recruitment process, we were gathering a group of experts to join the school. While the legal background of our SICSS and the previous experience from 2022 and 2024 (photos, stories, expert) were a strong foundation of our invitations we faced a number of challenges. The first issue concerned the timing of sending invitations. In previous editions, waiting for the final decision on additional funding for experts’ remuneration caused us to lose valuable opportunities to secure their availability. We learned from this experience and, in 2025, began reaching out much earlier, even before all funding was formally confirmed. As a result, we were able to invite and secure the participation of several outstanding experts from around the world, significantly enriching the academic quality of this year’s program. Some were surprised to be offered remuneration and were prepared to take part in the event pro bono as well. We took care of the remuneration of the experts in order to, on the one hand, ensure an appropriate level of their involvement (e.g., preparation of a dedicated presentation, pre-event online meeting to agree upon the content etc.) and, on the other hand, contribute to the professional image of the SICSS brand, as well as of our Faculty.
To the selected group that confirmed their participation, we provided a range of resources. This year, we decided to send a comprehensive brochure that gathered all essential information in one place — including how much time participants should reserve in their personal schedules, instructions on joining our Slack workspace, and guidance on preparing for the group projects. We also added practical details such as directions on how to reach the university campus, information about daily meals and scheduled coffee breaks, as well as suggestions for things to do in the area during their free time. This helped participants feel better prepared and more comfortable before their arrival. This year we did not do a dedicated boot-camp, as it was not popular last time, and besides that we switched to a low-code approach towards programming (using pre-made code on Google Colab, using LLMs as a help in coding). It seems that the practical aspect of the school was not fully obvious to some participants, who were (positively) surprised by the number of exercises and programming tasks. However, this does not necessarily mean they were well-prepared to handle them with ease — a conclusion similar to the one drawn in 2022 and 2024. In the 2025 edition, although many participants were not familiar with coding, we designed parallel tracks and activities for both coding and non-coding participants, ensuring that everyone had the opportunity to contribute meaningfully and learn something new. At the same time, we also had participants with very advanced technical skills, for whom some programming exercises may have been too easy. Striking the right balance remains challenging, as it is difficult to design activities that perfectly satisfy participants at every skill level. Still, the overall results show that a well-planned structure allows both beginners and experienced participants to grow, collaborate, and produce high-quality work.
We encountered no excessive correspondence regarding organizational matters. The only recurring comment from participants concerned the timing of the school — several of them indicated that they would prefer the Institute to take place in September rather than in June. As noted earlier, we received a limited response from some of the candidates accepted to the Institute. Out of 65 applicants, 25 were selected, yet only 16 ultimately took part in the program. In such cases, it would be helpful to follow up and learn about their reasons — whether personal circumstances or a mismatch with the Institute’s agenda. Unfortunately, this situation also meant that potentially motivated candidates lost their opportunity to participate simply because the available spots had already been allocated. This confirmed once again that no-shows can create unnecessary losses for applicants who were genuinely interested and would have fully benefited from the program.
The first week followed a structure similar to previous editions, but with several improvements based on past experience and the shift to an in-person format. We officially began on Monday with a welcome coffee and an informal meet-and-greet session, which replaced the virtual Sunday gathering used in previous online editions. Instead of a standard introductory round, participants engaged in an interactive quiz about their backgrounds and motivations, followed by short speed-meeting rotations that helped participants connect across academic disciplines early on.
Throughout the first week, we maintained the traditional SICSS format of combining lectures with hands-on guided-project sessions. Each day included a lecture slot, coffee break, lunch, and extended group-work blocks conducted in smaller rooms. Because the school was held fully on site this year, we dedicated specific spaces for teamwork and ensured participants had continuous access to refreshments, including coffee breaks and daily lunches organized at the university. This significantly strengthened the group dynamic and created a more collaborative environment than in previous online editions.
Lectures and expert sessions
The lecture series featured several distinguished guests. We hosted:
All lectures were intentionally programming-free, focusing instead on methodological foundations, ethical considerations, and broader conceptual frameworks. This balance helped participants alternate between intellectually intense conceptual sessions and more technical hands-on blocks.
Workshops and guided projects
The technical workshops—run daily after the morning break—were structured around small-group guided projects. Participants worked in fixed rooms assigned to specific teaching assistants, which supported consistency and individualized feedback.
Because participants arrived with highly varied levels of technical preparation, we designed exercises with two parallel tracks:
This structure proved highly effective. Beginners were able to make steady progress with the support of teaching assistants, while more advanced participants often completed tasks more quickly and were offered optional extended exercises. Even so, it remains challenging to completely calibrate the difficulty level so that all participants feel equally challenged, and this is something we will continue refining.
The pedagogical focus of the week covered topics central to our theme: how social polarization influences law and legal decision-making. Across lectures and workshops, we introduced participants to:
As in previous years, NLP and LLM-based content proved most engaging. Nearly all participants chose to rely on these techniques in their second-week projects, confirming the relevance of emphasizing these methods throughout the workshops.
This year, the on-site format strengthened community building. Lunches—served daily—naturally turned into brown-bag-style discussions, enabling participants to begin brainstorming project ideas early in the week.
A central component of the first week were the guided projects, conducted under the close supervision of the teaching assistants. Each TA designed a dedicated mini-project, which participants worked on throughout the week in smaller, stable groups. This structure ensured that every participant—regardless of prior technical background—could follow a clear learning path, receive individualized support, and gradually develop practical skills needed for the second-week group projects. The close TA involvement also helped maintain a consistent pace of learning and created a supportive environment where participants could ask questions freely and troubleshoot challenges in real time.
In addition to the academic program, this year’s first week placed particular emphasis on social activities that helped participants build stronger connections outside the classroom. Mid-week, we organized an evening integration outing that allowed the group to unwind and get to know one another in a more informal setting. On Saturday, a group of approximately fifteen participants joined a boat trip on the Warta River, which turned out to be one of the highlights of the week and greatly strengthened the sense of community. A smaller group also had the opportunity to attend the Pyrkon festival—an activity awarded to the winners of our competition—which offered a unique cultural experience and further contributed to team bonding. These social events played a meaningful role in fostering collaboration, improving group dynamics, and creating a friendly atmosphere that carried over into the second-week group projects.
The second week of SICSS-AMU/Law was devoted primarily to unguided projects, giving participants full independence in choosing their topics, designing methodologies, and organizing their workflow. After the structured, TA-led guided projects of week 1, this shift allowed students to explore their own ideas and apply newly acquired skills in a realistic research setting. The teaching assistants remained available for technical support, but the conceptual direction and execution were left entirely to the groups—an approach that once again proved to be one of the most effective pedagogical features of SICSS.
Tuesday was dedicated to the KJM Conference, which offered participants a unique opportunity to expand their interdisciplinary horizons and engage with ongoing research in neurotechnology and responsible innovation. The conference featured:
Participants consistently emphasized that attending the Conference broadened their perspectives, provided inspiration for their own projects, and complemented the legal-tech focus of the Summer Institute.
Throughout the week, we also welcomed several invited guests whose contributions added depth and practical relevance:
These contributions helped participants better understand the broader relevance of computational social science beyond the academic setting and provided them with concrete inspirations for their Week 2 work. These talks enriched the second week and added substantial value to the project development process.
From Wednesday to Friday, participants focused intensively on their unguided projects. They organized themselves into small groups, independently selected topics, explored datasets, and iteratively adjusted their methods. Two participants chose to work individually, which also fit well within the flexible structure of the week.
The final outcomes of the Unguided Project Workshop were:
These projects showed substantial progress within just a few days of unsupervised work and highlighted how effectively participants transitioned from structured exercises to independent research.
Despite the intensive nature of the week, we maintained opportunities for rest and socialization.
Participants attended a guided tour of the Lech Poznań Stadium. These events helped balance the workload, strengthened group cohesion, and facilitated informal exchanges across project teams.
The week concluded with final presentations held in the conference room at Lech Stadium, which added a professional atmosphere to the closing ceremony. Each group presented the motivation, methodological pipeline, datasets, and tentative findings of their project. The diversity of topics and the quality of initial results demonstrated both the creativity and determination of the participants.
The Institute concluded in a warm, friendly atmosphere, with participants openly expressing a sense of achievement—especially those who made remarkable progress in coding and data-analysis skills over the two weeks. Many emphasized how surprising and unique it felt to take part in such an intensive, high-quality programme completely free of charge, noting that comparable international training events are often difficult to access due to cost. This made the positive reception of the 2025 edition even more meaningful.
After the end of the school, participants continued to stay in touch with us and with each other. Several groups expressed a strong commitment to further developing the projects they started during the Institute, and we remain in active contact with them to support their next steps. At least two teams have already signalled clear plans for preparing publishable research outputs, which we see as one of the most promising outcomes of this year’s edition.
To evaluate the overall performance of the programme, we conducted short anonymous surveys throughout the Institute. The feedback we received consistently highlighted two aspects as the greatest strengths of SICSS-AMU/Law:
Several participants shared heartfelt messages of appreciation, highlighting the organizers’ dedication and the unusually supportive learning environment. Many mentioned that they did not expect such a comprehensive and professionally prepared event—especially in its first fully in-person edition—and expressed gratitude for the care and attention put into every aspect of the programme.
We are convinced that the success of the 2025 edition was possible not only because of the support we received from partner institutions and the impressive engagement of participants, but also thanks to a committed organizing team that worked together closely, looked after one another, and consistently went beyond formal responsibilities. The outcomes of this year’s Institute suggest a strong foundation for future editions and confirm that offering such an intensive, interdisciplinary programme in a fully in-person format was a step in the right direction.