Data Science for Asylum Legal Landscaping (DATA4ALL)
Leveraging large-scale decision data in the Nordics, DATA4ALL pioneers a new research agenda combining data science and migration law to understand the significant outcome variations in asylum decisions within and across countries.
Data Science for Asylum Legal Landscaping (DATA4ALL) is a four-year project (2020-2024) funded by a cross-cutting grant from the DATA+ pool administered by the Rector of University of Copenhagen. DATA4ALL will compare large-scale decision data from the Nordic countries and use natural language processing (NLP), sentiment analysis, machine learning, and process mining to unpack and provide a unique understanding of this phenomenon. The project further draws on critical data studies to engage decision-makers themselves, raise questions to the data and promote data literacy and ethics among both scholars and practitioners.
The project is organised around three tracks:
Track 1 examines static factors impacting recognition rates year-to-year and across countries. Key factors include applicant age, gender, level of education, ethnicity, religion, tribe and local origin/residence. It uses natural language processing (NLP), machine learning and clustering to enable identification of relevant correlations with success rate and meaningful clusters.
Track 2 draws on participatory methods and critical data studies approaches to raise questions to the data based on the socio-technical practices of decision-makers. The combination of NLP and participatory methods allows us to move beyond “obvious” application areas, enables grounded sensemaking of data as well as co-development of practical approaches to a responsible data science in this area.
Track 3 adds a temporal dimension using event log construction to understand decision-making workflow and sentiment analysis to determine if attitudes towards applicants change during the process or over time. Techniques from time series analysis and process mining will be used to analyze whether recognition rates are impacted by procedural changes. This final stage of the project will enable us to offer concrete recommendations for decision-support technologies.
The project is organised in three work packages grounded in the legal, machine learning, and critical data science perspectives, respectively headed by Thomas Gammeltoft-Hansen, Tijs Slaats, and Naja L. Holten Møller.
The project will be implemented as an interdisciplinary collaboration between the Faculty of Law and the Department of Computer Science, University of Copenhagen and with partners at the University of Oslo (Norway) and the University of Uppsala (Sweden).
The project further engages with key stakeholders in the field of Nordic asylum law, including Danish Refugee Council, Danish Refugee Appeals Board, Norwegian Organisation for Asylum Seekers, Swedish Refugee Law Center and UNHCR Northern Europe.
|Byrne, William Hamilton
|Professor with special responsibilities
(PI, NordASIL; PI, DATA4ALL)
|Hildebrandt, Thomas Troels
|Høgenhaug, Anna Højberg
|Møller, Naja L. Holten
Assistant professor, tenure track
|Nielsen, Trine Rask
|Olsen, Henrik Palmer
|Professor in Jurisprudence
|Associate Professor, tenure track
(PI, DATA4ALL; WP-leader, NordASIL)
Data Science for Asylum Legal Landscaping (DATA4ALL) has received a four year funding from the DATA+ pool administered by the Rector of University of Copenhagen.
Period: 1 september 2020 – 31 August 2024.