The following plenary sessions are planned for International Data Week 2023. Preparations by the Programme Committee are underway and confirmed speakers will be announced soon.
Inclusivity in Open Science while advancing research assessment and career pathway impact
SESSION OVERVIEW: The UNESCO Recommendation on Open Science. In adopting the Recommendation in November 2021, 193 countries have agreed to abide by common standards for open science. In May 2013, the San Francisco Declaration on Research Assessment was published, and DORA was established. DORA’s vision is to advance practical and robust approaches to research assessment globally and across all scholarly disciplines. Furthermore, the OECD is working with member and non-member economies to review policies to promote open science and to assess their impact on research and innovation. These, and many other, global milestones are fundamental to achieving the open science vision.
But how can we ensure that in defining and developing open science solutions and standards, the principle of inclusivity is captured by the global movement to improve the ways in which research and scientific outputs are evaluated? How can we advance Research Assessment with an increased focus on the quality of research outputs rather than quantity, and by fit-for-purpose use of diversified indicators and processes that forego the use of journal-based metrics such as the journal impact factor?
The collective, global aim is to support research practices that are more transparent, collaborative and inclusive are subject to more effective peer review, increased scrutiny and critique which in turn increases the verifiability and reproducibility of the science produced. This ultimately leads to better science, more trust in science and more relevant and positive impacts of science on society.
SESSION FOCUS: This session will focus on the challenges and urgency of inclusivity[1] as part of the global research assessment reform. What are the needs of indigenous groups and other minority groups in this context? What needs to be addressed? Who can and should support this change? What is the role of the global research and science communities in this change?
Data and global challenges: data, science, trust and policy
The major global human, societal, and scientific challenges of our age are fundamentally interdisciplinary and related to all sectors of society. These challenges can only be addressed through the close collaboration of science, civil society, and government using cross-domain and multi-stakeholder research that seeks to understand complex systems, including through machine-assisted analysis at scale.
The Open Science, the FAIR Principles and good data stewardship are central to the success of such research. In combination, as part of good scientific practice, these things enable the efficient and reliable collection and processing of data, and they allow provenance and processing to be recorded in a transparent and robust manner. Taking into account methodological issues, inaccuracies and error bars, these data provide the raw material for science to create knowledge and provide evidence to guide policy and action.
This session will seek to explore these processes at a global scale, including the challenges of creating ‘global observing systems’ to gather the data we need; the need for transparency and traceability through the processes of data integration, modelling and the application of complex, machine-assisted analysis; the hazards of turning science into action, when operating at the science-policy interface; and above all the need to engender trust throughout these processes.
- Global observing systems.
- Transparency and traceability.
- Complex systems and evidence
- Science and Policy
Ethics, data science, and AI in dialogue with one another and society
This session examines the developing relationship between ethics, data science, and artificial intelligence in relationship to one another and to society. Data science brings with it not only new understandings of our work, but (arguably) it opens new worlds for exploration and new pathways for science. Artificial Intelligence similarly transforms science (and society) fundamentally by revising our scientific methodologies while also providing new pathways for decision-making and narration. This session opens a dialogue on the fundamental frameworks of ethics in the data sciences and artificial intelligence, focusing on the human dimensions of the technological impacts of our digitalised societies on fundamental human rights, social values, and legal frameworks in our differentiated yet shared human conditions.
Take away / outcomes
The session helps to frame a principal theme of this conference: the need to promote fundamental human rights and democratic values through our digitalised societies. This session aims to contribute of a multi-disciplinary understanding as to how ethics and data science can enrich one another and, so doing, contribute to addressing the fundamental challenges of our human condition today.
Spatial Data Science: Geographic Context Matters
Understanding context factors in data science is key for drawing reliable and robust conclusions. As such, geographic space is one of the most strongly defining context factor there is, providing the possibility to reveal geospatial relationships, structures and dependencies through analysing digital geospatial data. Consequently, we can meaningfully interpret data-scientific results against its geographic characteristics and context factors.
Although a large majority of datasets contain spatial information, traditional data science methods have been using a variety of attributes to examine real-world phenomena, but mostly neglecting the spatial nature of the phenomenon under investigation. Yet, a wide variety of application domains and data science scenarios require an explicit spatial view or benefit from it to gain additional insights. For instance, analysing and visualising spatial relationships and structures is vital for efficiently extracting in-depth information for decision support in areas like disaster management, public health, urban analytics, humanitarian aid, energy systems research or ecology.
Recent efforts have increasingly promoted this explicit spatial view through combining legacy data science methods with spatial analysis paradigms. However, the major research challenge we are currently facing in data science and machine learning algorithms is the lack of handling geospatial data. Previous approaches have engineered features based on derived geospatial metrics like distance or density, but are not able to handle geospatial coordinates as features per se.
Thus, inter- and multidisciplinary approaches are required to 1.) tackle the methodological challenges sketched out above, and 2.) to foster improved communication between scientific disciplines and application domains for generating a better understanding of spatial processes. This, in turn, leads to significantly enhanced decision-making through a more holistic view and a more comprehensive information base.
The session will cover the following topics, but is not strictly limited to them:
- Spatial data science in disaster management, public health, urban management, humanitarian aid, energy systems research, ecology, a.o.
- Multimodal data analysis (geospatial, temporal, semantic, image, etc.)
- Spatially explicit machine learning and geoAI
- Geographic question answering
- Integrated spatio-temporal analysis
- Integrating domain knowledge with spatial data science methods
- Analysis vs. modelling and simulation of geospatial processes
- Digital Twins
- Location privacy
- Interoperability and transferability