Speaker
Description
Data-driven Spatial Decision Support Systems (SDSS) hold significant potential for addressing various municipal tasks, including financial planning, the siting of social facilities, and crisis management. It furthermore suggests objective assessment and evaluation of criteria, but this is only achievable when the underlying data is of high quality and free from biases, and when the algorithm is carefully designed at all critical points (Sugumaran and DeGroote, 2010).
Municipal registry data is an often used and reliable source with mostly standardized attributes for population data. This data is diligently maintained by local authorities, yet it is not entirely free of certain biases. One such bias is the so called undercoverage bias, which signifies that not all inhabitants of a municipality are recorded in the registry. This often relates to marginalized groups, such as homeless individuals, seasonal workers, or people without legal residence status (Bundesministerium für Arbeit und Soziales, 2022). Special provisions exist in some cases, for example, NATO personnel in Germany are exempt from registration requirements (Bundesministerium der Justiz, 2020). Quantifying undercoverage bias in registry data is challenging to impossible as certain population groups such as the homeless can only be estimated, even by experts. However, the absence of certain groups certainly has an impact on the presentation of a population and its attributes in SDSS and this may have significant implications for decision-making processes. It is therefore desirable to transparently display known information about biases and unknown aspects in used databases rather than simply accepting deficits in data quality and not communicating them to users.
This paper investigates the extent to which undercoverage bias in registry data impacts the accuracy and reliability of SDSS visualizations. It also examines how data aggregation at different spatial levels might amplify or mitigate these effects, depending on the spatial distribution of missing data. Common visual representations, such as bar charts, pie charts, and choropleth maps, are analyzed to illustrate these dynamics.
While significant impacts from undercoverage bias may not always be evident, even the absence of effects can provide valuable insights into the resilience of SDSS against data deficits. The study underscores the importance of transparently communicating both known limitations and uncertainties in underlying datasets to SDSS users, enabling informed decision-making processes and encouraging critical thinking among users regarding these systems. In future research, this work will serve as a foundation to explore methods for effectively conveying uncertainties and gaps in knowledge to users of SDSS. By developing communication strategies and visualization techniques, the aim is to ensure that users can better understand and navigate the implications of data deficits in spatial decision-making processes.
References
Bundesministerium für Arbeit und Soziales (2022), Ausmaß und Struktur von Wohnungslosigkeit - Der Wohnungslosenbericht 2022 des Bundesministeriums für Arbeit und Soziales. [Online] available at: https://www.bmwsb.bund.de/SharedDocs/downloads/Webs/BMWSB/DE/veroeffentlichungen/pm-kurzmeldung/wohnungslosenbericht-2022.pdf
Sugumaran, Ramanathan, and DeGroote, John. Spatial Decision Support Systems: Principles and Practices. Baton Rouge: Taylor & Francis, 2010. Print. Pp. 442 ff.
Bundesministerium der Justiz (2020). Streitkräfteaufenthaltsgesetz. [Online] available at: https://www.gesetze-im-internet.de/skaufg/
| Keywords | Spatial Decision Support Systems; Registry Data; Bias; Visualizations; Minority Groups |
|---|---|
| Best Congress Paper Award | Yes |
