FIN-CLARIAH

In 2024-2025, the FIN-CLARIAH infrastructure project continued the work of the previous project periods. The aim was to expand support for humanities and social science research by developing digital practices and tools for researchers.

The University of Jyväskylä's work in 2024-2025 focused on the following themes:

D3.3.1 Automated metadata of archival data from NAF

This work package is carried out in collaboration with the National Archives of Finland. The aim is to create ways to automatically enrich archival data with metadata. During the project period, research will focus on the use of multimodal datasets and models for the classification and enrichment of diverse archival data.

Team: Antero Holmila (work package leader), JYU, Ida Toivanen, JYU, Venla Poso, JYU, Tanja Välisalo (KA)

D3.3.3 Machine-learning-based enrichment of social media (Twitch) / D4.1.1 Analysis of video stream interactions with AI solutions

The work package focuses on machine learning-based enrichment of livestream data and their chat data, and AI-based interaction analysis of livestreams. Tools for data collection and analysis will be developed for use by researchers.

Team: Raine Koskimaa, Work Package Leader, JYU, Jari Lindroos, JYU, Tanja Välisalo, JYU, Jaakko Peltonen, University of Tampere.

Finnish and Finland-Swedish Sign Language corpora - WP1.3 Video processing and annotation / WP2.1 Personal and Copyrighted Research Data

The work at the University of Jyväskylä is related to two work packages within FIN-CLARIAH. The work at JYU is related to the publication of three sign language corpora (two Finnish sign language corpora and one Finnish-Swedish sign language corpus). Juhana Salonen has been responsible for the implementation of the work on the JYU side.

https://converis.jyu.fi/converis/portal/detail/ResearchDataset/177402998?lang=fi_FI

https://converis.jyu.fi/converis/portal/detail/ResearchDataset/160087813?lang=fi_FI

https://www.kielipankki.fi/aineistot/cfsts/

In the first common development project in 2022–2023, the FIN-CLARIAH components aimed to significantly broaden their mutual scope of digital SSH infrastructural support by consolidating and enhancing their resources in three major data-oriented directions:

1. To reach beyond processing of spoken standard Finnish into colloquial speech 
2. To cater to a broad range of SSH research needs for processing unstructured text 
3. To facilitate research based on metadata 

Github: JYU-digihum

Project lead: Professor Jari Ojala

FIN-CLARIAH project included eight Finnish universities, National Library, National Archives, CSC  – IT Center for Science Ltd. and Institute for the Languages of Finland.

University of Jyväskylä (JYU) lead the the work in following working packages in 2022–2023:

WP 2.5 Solutions for better use of language learner performances in research

The aim of the WP2.5 is to provide infrastructure for better use of language learner performances in research. Current language tools for research do not fully adapt to learner language with its unique features. Accurate assessment of language proficiency level can be crucial as it can affect e.g. immigrants’ possibilities to apply for citizenship. To help reduce the human error, an automatic assessment tool is needed. In this project, we develop a first version of an AI-based tool to evaluate language learner proficiency in Finnish language.

Team: Ari Huhta, lead, JYU, Jenny Tarvainen, JYU, Sirkku Kronholm, JYU (project alumni)

WP 3.2 AI solutions to better use National Archives mass digitisation services

The aim is to develope AI based solutions to help researchers in various fields to effectively use massive digitized datasets from the National Archives of Finland. In this work package, the usability and possibilities for analysis are improved by developing an AI based model for Named Entity Recognition (NER). The new model understands OCR'd archival data as well as state administrative language better than previous models, and also includes two brand new entities: business identity codes and journal numbers.

Team: Antero Holmila, lead, JYU, Venla Poso, JYU, Ida Toivanen, JYU

WP 3.3 AI-based solutions to better use textual qualitative survey data

The aim is to develop tools for analysing open-text survey responses written in Finnish. In many cases, open responses in survey data are not properly utilized due to lack of resources, especially easy-to-use coding packages tailored for Finnish language. To support researchers conducting qualitative analysis, we create: (1) Concept Network - a tool for identifying key concepts and their connections, and (2) an AI based clustering tool for grouping together semantically similar open responses.

Team: Krista Lagus, lead, University of Helsinki, Maria Valaste, UH, Sakari Taipale, JYU, Tapio Litmanen, JYU, Tomi Oinas, JYU, Ida Toivanen, JYU, Joni Oksanen, HY, Jani-Matti Tirkkonen, University of Eastern Finland, Jaakko Peltonen, Tampere University

WP 3.4 Developing analysis methods for real-time chats in game play stream

The aim is to develop analysis methods for real-time chat in gameplay streams. The proliferation of livestream video has impacted audience engagement with media and chat feature is a central part of this phenomenon. Chat analysis has been challenging due to the massive amount of messages as well as internet language. For collecting and analysing chat data we develop tools that enable analysing the dynamics of chat interaction and the content of chat messages

Team: Raine Koskimaa, lead, JYU, Jari Lindroos, JYU, Tanja Välisalo, JYU, Jaakko Peltonen, Tampere University

• Lindroos, J., Toivanen, I., Peltonen, J., Välisalo, T., Koskimaa, R., & Äyrämö, S. (2025). Participant profiling on Twitch based on chat activity and message content. In P. D'Souza, H. Tang, N. Xi, & J. Hamari (Eds.), GamiFIN 2025: Proceedings of the 9th International GamiFIN Conference (4012, pp. 267-282). RWTH Aachen. CEUR Workshop Proceedings. https://ceur-ws.org/Vol-4012/paper18.pdf
• Tarvainen, J., Toivanen, I., & Huhta, A. (2025). Automatic language proficiency assessment of written texts: Training a CEFR classifier in L2 Finnish. Studies in Language Assessment, 14(2), 58-90. https://doi.org/10.58379/YWAV5140
• Toivanen, I., Poso, V., Lipsanen, M., & Välisalo, T. (2025). Developing named-entity recognition for state authority archives. In O. Holownia, & E. S. Sigurðarson (Eds.), DHNB2024 Conference Post-Proceedings (7). University of Oslo Library. Digital Humanities in the Nordic and Baltic Countries Publications. https://doi.org/10.5617/dhnbpub.12262
• Lindroos, J., Peltonen, J., Välisalo, T., Koskimaa, R., & Toivanen, I. (2025). From PogChamps to Insights : Detecting Original Content in Twitch Chat. In T. X. Bui (Eds.), HICSS 2025: Proceedings of the 58th Annual Hawaii International Conference on System Sciences, (pp. 2542-2551). University of Hawaiʻi at Mānoa. Proceedings of the Annual Hawaii International Conference on System Sciences. https://hdl.handle.net/10125/109148
• Poso, V., Lipsanen, M., Toivanen, I., & Välisalo, T. (2024). Making Sense of Bureaucratic Documents : Named Entity Recognition for State Authority Archives. In Archiving 2024 Final Program and Proceedings (pp. 6-10). Society for Imaging Science & Technology. Archiving, 21. https://doi.org/10.2352/issn.2168-3204.2024.21.1.2
• Poso, V., Välisalo, T., Toivanen, I., Holmila, A., & Ojala, J. (2023). Untapped data resources : Applying NER for historical archival records of state authorities. In A. Rockenberger, J. Tiemann, & S. Gilbert (Eds.), DHNB2023 Conference Proceedings (5, pp. 55-69). University of Oslo Library. Digital Humanities in the Nordic and Baltic Countries Publications. https://doi.org/10.5617/dhnbpub.10650
• Toivanen, I., Räsänen, V., Lindroos, J., Oinas, T., & Taipale, S. (2022). Implementing sentiment analysis to an open-ended questionnaire : Case study of digitalization in elderly care during COVID-19. In J. Domenech, & M. R. Vicent (Eds.), CARMA 2022 : 4th International Conference on Advanced Research Methods and Analytics (pp. 205-212). Editorial de la Universitat Politècnica de València. International Conference of Advanced Research Methods and Analytics. https://doi.org/10.4995/CARMA2022.2022.15089
• Koskimaa, R., Välisalo, T., & Lindroos, J. (2022). Understanding Twitch Esports Communities through Livestream Chat Analysis. In #ERNC22 Esports Research Network Conference : Sustainable Esports in the Digital Society : Book of Abstracts (pp. 139-141). Esports Research Network. https://esportsresearch.net/wp-content/uploads/2023/01/ERNC22_Jonkoping_Book_of_Abstracts.pdf
• Toivanen, I., Lindroos, J., Räsänen, V., & Taipale, S. (2022). Dealing with a small amount of data : developing Finnish sentiment analysis. In 2022 BESC : 9th International Conference on Behavioural and Social Computing. IEEE. https://doi.org/10.1109/besc57393.2022.9995536