Resources

Tools to make sense of web data

This resource consists of two tools: one to classify toxic data in Finnish (e.g., insults, obscene language) from datasets retrieved from social media platforms; and another to identify registers (genres, e.g., reviews, interviews, news reports) from web content in diverse languages.

Toxicity classifier: https://github.com/TurkuNLP/toxicity-classifier
Multilingual modeling of web registers: https://github.com/TurkuNLP/multilingual-register-labeling

Resource developed by the TurkuNLP / University of Turku in partnership with the CSC.
Guidance can be found in the websites of the resources.

Twitch Chat Collector & Analysis Tool

This resource collects chat data from the live stream service Twitch and YouTube. Thanks to this resource, researchers will be able to retrieve and analyze larger samples of chat data from the livestream services Twitch and YouTube.

The tools sidebar contains multiple ways to collect data, but also sections for chat content classification based on machine learning and video clip analysis based on Multimodal Large Language Models.

Resource developed by the University of Jyväskylä with collaboration from Tampere University.
Guidance can be found in the website of the resource.

Document Understanding Tools

The archival data team, consisting of Venla and Ida, has worked on producing tools for document understanding. This refers to various kinds of processing of documents, such as named entity recognition and document type classification.

Named entity recognition (UI): https://arkkiivi.fi/ 

Named entity recognition (Huggingface): https://huggingface.co/Kansallisarkisto/finbert-ner 

Document type classification https://huggingface.co/jyu-digihum/findoctype 

Most of the tool development has been conducted in collaboration with the National Archives of Finland.

L2 Finnish model

L2 Finnish model is a classification model trained with CEFR annotated data containing fictional and non-fictional texts written by Finnish as a second language (L2) speakers. With the model you can classify texts into the following CEFR classes: A1, A2, B1, B2, and C.

NTS: Nordic Tweet Stream

This resource makes Twitter/X data available for researchers. Altogether, it contains nearly 74 million messages from hundreds of thousands of user accounts from the five Nordic countries. The NTS data cover the period between January 2013 and May 2023 and were collected using the Academic API, which is now closed. The NTS comes with an easy-to-use graphic interface that supports quick data access. It is possible for instance to study public discourses and sentiment concerning events in recent history. Researchers will be able to search, subset, visualize, and download data.

Access to resource: https://nordictweetstream.fi/

Resource developed by the University of Eastern Finland, in collaboration with Linnaeus University.
Contact information and guidance can be found in the website of the resource.

Sampo UI builder

This resource provides a framework for building customizable and responsive user interfaces for semantic portals without the necessity of having broad coding skill.

Sampo UI: https://seco.cs.aalto.fi/tools/sampo-ui/
Tutorial: https://seco.cs.aalto.fi/tools/sampo-ui/Sampo-UI-tutorial.pdf
An example of semantic portal created with this resource is ParliamentSampo: https://parlamenttisampo.fi/ . In this portal, it is possible to test its functionality to study parliamentary speeches; as well as some examples of queries that can be addressed using the portal.

Resource developed by Aalto University in partnership with the University of Turku and the University of Helsinki.

Text Network Tools for Parliamentary Data

This resource provides tools based on network analysis for the analysis of political text. With these tools, researchers will be able, for example, to analyze keyword embeddings of the FinParl corpus and identify how phrases or longer text passages are re-used over time in he MPs plenary debates of the Finnish parliament.

KWIC keyword tool for FinParl corpus: http://finparl-01.utu.fi/apps/KWIC/
TNA tool for the analysis of speeches of Finnish MPs: http://finparl-01.utu.fi/apps/TNA

Resource developed by the University of Turku in partnership with Aalto University. Collaborators: the University of Jyväskylä.

Finnsurveytext

This resource provides a set of easy-to-use tools for conducting qualitative analysis on survey responses in Finnish. Thanks to this resource, researchers will be able to better understand data retrieved from open-ended questions.
CRAN webpage: https://CRAN.R-project.org/package=finnsurveytext
Guidance can be found in the website of the resource.

Finnish Forum Scrapers

This is an application for scraping comment-data from Finnish resources with high user traffic.

Access: https://github.com/uh-dcm/finnish-forum-scrapers

Historical Newspapers in the CSC Supercomputing Environment

This resource allows to download copyright-free materials from the National Library of Finland through the CSC.

Access the resource: https://github.com/CSCfi/kielipankki-nlf-harvester
Technical documentation: https://urn.fi/urn:nbn:fi:lb-202311261

Resource developed by the National Library of Finland in partnership with the CSC, University of Helsinki and University of Turku. Collaborators: National Archives of Finland and University of Jyväskylä.

Harmonized Finnish National Bibliography

This resource provides a harmonized version of the Finnish national bibliography (Fennica) dataset as well as the code used for cleaning, enriching and automatically generating reports on the data. Thanks to this resource, researchers will be able to extract bibliographic metadata for large scale statistical analysis.

Access to resource: https://fennica-fennica.2.rahtiapp.fi/
Code use to harmonize metadata: https://github.com/fennicahub/fennica

Information and guidance can be found in the webiste of the resource.
This resource has been developed by the University of Turku in partnership with the University of Helsinki. Collaborators: National Library of Finland, University of Jyväskylä.

Tool to evaluate biases and errors

This resource provides tools for subsetting and evaluating datasets that have not originally been created for research. Thanks to this resource, researchers will be able to robustly explore large datasets, examine their representativeness, and extract the subset they are interested in.

End-user interface links and usage instructions for centrally indexed datasets: https://github.com/hsci-r/elasticsearch-openshift/blob/main/documentation/exported_query.md
Technical documentation enabling people to set up their own instances for their own datasets: https://github.com/hsci-r/elasticsearch-openshift

Resource developed by the University of Helsinki (ARTS) in partnership with the CSC.

Forensic Linguistics Corpus and Search Interface C.R.I.M.E.

This resource is a structured, searchable corpus comprising audio and ASR-generated transcripts from investigative interviews, courtroom interactions, and related media.

Access the database: https://forensic.corpora.li
Access the static dataset: https://doi.org/10.7910/DVN/MLMB6E

Additional information (user guide, proceedings article) are linked on the websites.

This resource has been developed by Steven Coats, University of Oulu

Automated Automated Harmonisation and Enrichment of Metadata

This resource provides R packages for collecting and enriching of Finnish cultural heritage metadata. finna R package is for collecting cultural metadata using the Finna API and the second is finto R package for enriching the metadata using the Finto API from the finto service. geofi R package is for Geospatial analysis and visualization of metadata. These tools are designed to offer easy access, geospatial analysis and visualization of metadata for cultural heritage researchers.

Finna R package: https://github.com/fennicahub/finna
Finto R package: https://github.com/fennicahub/finto
Geofi R package: https://github.com/rOpenGov/geofi

Information and guidance can be found in the webistes of the resources.

This resource has been develeoped by the University of Turku. Collaborators: National Library of Finland.

DARIAH-FI Zotero library

A public directory of publications (research articles, conference proceedings, data publications) that point at, explain or introduce use cases for the infrastructures developed by the DARIAH-FI partners for the FIN-CLARIAH project.

User Experience Questionnaire

UX questionnaire developed within DARIAH-FI to test and evaluate tools, datasets or workflows developed for the project. The questionnaire was created and updated in several phases between 2022-2023 from a literature review, semi-structured interviews, and tests with end-users.