Bibtex
Cite as text
@Select Types{,
Journal = "Band-1",
Title= "Hybrid Intelligence with Commonality Plots: A First Aid Kit for Domain Experts and a Translation Device for Data Scientists",
Author= "Nikolas Stege, Michael H. Breitner",
Doi= "https://doi.org/10.30844/wi_2020_c7-stege",
Abstract= "There is a large gap between domain experts capable to identify business needs and data scientists who use insight producing algorithms, but often fail to connect these to the bigger picture. A major challenge for companies and organizations is to integrate practical data science into existing teams and workflows. We are driven by the assumption that efficient data science requires cross-disciplinary teams able to communicate. We present a methodology that enables domain experts and data scientists to analyze and discuss findings and implications together. Motivated by a typical problem from auditing we introduce a visualization method that helps to detect unusual data in a subset and highlights potential areas for investigation. The method is a first aid kit applicable regardless whether unusual samples were detected by manual selection of domain experts or by algorithms applied by data scientists. An applicability check shows how the visualizations facilitate collaboration of both parties.
",
Keywords= "Commonality Plots, Domain Knowledge, Hybrid Intelligence, Visualization, Data Science
",
}
Nikolas Stege, Michael H. Breitner: Hybrid Intelligence with Commonality Plots: A First Aid Kit for Domain Experts and a Translation Device for Data Scientists. Online: https://doi.org/10.30844/wi_2020_c7-stege (Abgerufen 26.12.24)
Open Access
There is a large gap between domain experts capable to identify business needs and data scientists who use insight producing algorithms, but often fail to connect these to the bigger picture. A major challenge for companies and organizations is to integrate practical data science into existing teams and workflows. We are driven by the assumption that efficient data science requires cross-disciplinary teams able to communicate. We present a methodology that enables domain experts and data scientists to analyze and discuss findings and implications together. Motivated by a typical problem from auditing we introduce a visualization method that helps to detect unusual data in a subset and highlights potential areas for investigation. The method is a first aid kit applicable regardless whether unusual samples were detected by manual selection of domain experts or by algorithms applied by data scientists. An applicability check shows how the visualizations facilitate collaboration of both parties.
Commonality Plots, Domain Knowledge, Hybrid Intelligence, Visualization, Data Science