ExpertiseUpdated on 21 November 2023

Progressive Visual Analytics

Hans-Jörg Schulz

Associate Professor at Aarhus University

Aarhus, Denmark

About

Tired of reading pitches? Watch the video instead! https://youtu.be/XsdFYcDnUM4

Progressive Visual Analytics is a very recent field of research and innovation with a first research agenda appearing in 2024. In essence it means to chunk up large datasets and process and visualize them chunk-by-chunk. Think Progressive JPEG - only for integrated machine learning and data visualization.

Progressive Visual Analytics offers a variety of unique advantages over traditional all-at-once data analysis:

  • Reduced analysis times: work with in-progress results as if they were final

  • Computational steering towards subsets of the data space and the parameter space while the analysis is running

  • Added explainability / trust in the analysis results as one watches the results unfold

  • Elegant way to ensure data minimalism by only processing as many chunks as necessary to yield a result of given precision

  • Framework for Visual Analytics as a Service (pay per chunk)

  • Resilience mechanism to deal with data/software quality issues as chunks causing issues (division by zero, unhandled exceptions...) can merely be put aside and the analysis can carry on on the remaining chunks

Progressive Visual Analytics can provide a modern methodology to conduct interactive data analysis in distributed scenarios like fog computing and other smart IoT setups.

Please reach out, if your consortium can benefit from such an advanced, incremental human-in-the-loop analysis!

---
Own Publications on Progressive Visual Analytics:

Organisation

Aarhus University

University

Aarhus, Denmark

Similar opportunities