Result description
Research led by University of Konstanz have resulted in advances in the combination of data and feature space that bridges gaps between machine learning and human reasoning. Data analysis, visualization, and interaction facilities are combined in novel ways to enable analysts to interact with techniques such as unstructured data analysis, dimensionality reduction, clustering, and pattern mining to foster sense-making in the area of comparative data analysis. The techniques are combined and visualized in a fashion such that users without any or little expertise in data science are capable of exploiting these techniques and their visualizations in, for example, comparative case analysis.
(1) Aggregated Visualization of Off-Screen Elements
Navigation and spatial context is key for exploration of large data spaces. State-of-the-art techniques provide zoom in (drill-down) features, but the analyst loses spatial context. With aggregated visualization of off-screen elements, the spatial context can be maintained.
(2) Interactive Machine Learning for Crime Data Analysis
The analysis of crime data is mostly a manual task. Visualization and interaction techniques have been developed to allow analysts to seamlessly engage with machine learning and automated data analysis.
(3) Interactive Dimensionality-Reduction to foster Data Exploration and Sense Making
Typical data exploration systems require manual handling of data attributes. Interactive dimensionality-reduction does not require the analyst to select interesting attributes. It allows a comprehensive view on the dataset without any preconditions/filtering.
(4) Knowledge Generation Model for Visual analytics
Provides a terminology and process model and illustrates how uncertainties propagate though visual analytic systems and how user’s awareness and trust affects knowledge construction in a visual analytics context.
Addressing target audiences and expressing needs
- Business partners – SMEs, Entrepreneurs, Large Corporations
- Collaboration
- Fellowship to advance my/our research
Collaborators to advance the research and to apply them in technology development in areas outside of policing.
- International Organisations (ex. OECD, FAO, UN, etc.)
- Other Actors who can help us fulfil our market potential
- Research and Technology Organisations
- Academia/ Universities
R&D, Technology and Innovation aspects
The interaction and visualisation techniques have been reported in the scientific literature and will be incorporated into a commercial product (VALCRI spinoff).
Result submitted to Horizon Results Platform by MIDDLESEX UNIVERSITY HIGHER EDUCATION CORPORATION
