visual analytics for decision automation
This project aims at applying visual analytics techniques to address some well-identified challenges of Business Rules Management Systems. Those challenges involve providing business users with tools to better understand the structure and properties of the business rules system they design, and leveraging analytics tools to better anticipate emerging effects among massive automated decision systems. This project was started around 2007 and carried through until 2013 at the IBM France Lab. I worked with Frank van Ham and co-advised Bertjan Broeksema's PhD thesis at the IBM France Lab. Some published results:
- Visual Analysis of Multi‐Dimensional Categorical Data Sets B Broeksema, AC Telea, T Baudel. Computer Graphics Forum 32 (8), 158-169
- Decision exploration lab: A visual analytics solution for decision management B Broeksema, T Baudel, A Telea, P Crisafulli Visualization and Computer Graphics, IEEE Transactions on 19 (12), 1972-1981
- Capturing the Design Space of Sequential Space-Filling Layouts. Baudel, T & Broeksema, B, IEEE Transactions on Visualization and Computer Graphics, Volume: 18 , Issue: 12 Page(s): 2593 - 2602.
- At Bertjan's defense, with Jean-Daniel Fekete.
- Contextual feedback of rules proximity based upon co-occurence history in a collaborative rule editing system. Baudel, T. Van Ham, F. US Patent 8805771.
- Rule correlation to rules input attributes according to disparate distribution analysis. Baudel, T. Van Ham, F. US Patent 8825588.
ilog discovery
In the 80's, the spreadsheet revolutionized the world of data analytics. However, it's paradigm doesn't scale when available datasets grow by order of magnitudes. ILOG Discovery provides means to visualize and manipulate very large datasets interactively. It defines a formal visualization model called "data-linear visualizations", which can be compared to the relational model of databases, and leverages this model to allow interactive construction of visualizations, as well as means to reverse the model to provide means to edit data through the inverse operators.
- Browsing through an Information visualization design space. Scientific video.
- From information visualization to direct manipulation. Scientific video.
- ILOG Discovery is a visualization tool for analyzing data sets intuitively and communicating findings within an enterprise information system. It also serves as an interactive tool for browsing and editing databases, using a direct manipulation paradigm instead of a form-based query system.
- Data-linear visualizations. ILOG research report explaining the model behind Discovery's flexibility.
- Browsing through an information visualization design space. CHI 2004 demo.
- InfoCrible: édition interactive de visualisations compactes. Version française de la démonstration.
- Visualisations compactes: une approche déclarative pour la visualisation d'information. Version française de l'article.
algorithm visualization
A quick introduction to information visualization.