Datasaurus Dozen [Matejka et al. 2017]

• Jean-Daniel Fekete (INRIA) - jean-daniel.fekete@inria.fr
• Julien Tierny (CNRS) - julien.tierny@sorbonne-universite.fr

• Friday, 4:15pm - 7:15pm
• Room 1002 (Sophie Germain building)

As of 2023, this class (previously dealing with graphics and visualization) is completely focused on visual data analysis.

Jean-Daniel Fekete, a leading expert in data visualization, is joining the teaching team. Welcome Jean-Daniel!

Data analysis is a central challenge in today's numerical society. While many advanced statistical tools have been proposed to analyze complex data, these may fall short when it comes to unravel the unexpected, as illustrated in the above animation, with the dinosaur shape surprisingly popping up in the celebrated “Datasaurus dozen” distribution collection [Matejka et al. 2017]. From a basic statistical standpoint, all these distributions are identical: they share the same means, the same standard deviations, and the same correlation. However, they are arguably different.

This observation motivates the need for visual and interactive tools, helping users to investigate their data and to leverage the power of the human visual system to improve data interpretation.

Moreover, as shown in the “Datasaurus dozen” example, in many scenarios, data descriptors which are more visual in essence (e.g. dealing with shape and topology) can be more informative.

This class will describe how to integrate a visual aspect in the data analysis process, either by (1) creating data visualizations or by (2) exploiting visual descriptors, in particular the topological descriptors from Topological Data Analysis (e.g. persistence diagrams, Reeb graphs, Morse-Smale complexes, etc.).

The course will include 8 classes of 3 hours (each class will include itself a 1.5h lecture and a 1.5h hands-on), as well as 1 session of 3-hours where students will present research papers.

1. [15/09][Tierny] Global introduction, visualization basics & Practical hands-on with ParaView [Bring your laptop] Slides
2. [22/09][Fekete] Introduction: the Human Processor and the Grammar of Graphics & Practical hands-on with the GoG Slides
3. [29/09][Tierny] Introduction to topological methods and their applications to visual analysis & Practical hands-on with TTK [Bring your laptop] Slides
4. [06/10][Fekete] Visualization of tables & Practical hands-on with multidimensional visualization Slides
5. [13/10][Tierny] Towards more advanced descriptors (Reeb graphs, Morse-Smale complex) & Practical hands-on with TTK [Bring your laptop] Slides
6. [03/11][Fekete] Visualization of Graphs & Practical hands-on with graph visualization Slides
7. [10/11][Fekete] Interaction with visualization & Practical hands-on with interaction
8. [17/11][Tierny] Research seminar
9. [24/11] Written exam (1 hour, 10 short open questions, 1 A4 sheet of hand-written notes allowed)

• Computational Geometry
• Algorithms for Geometric Graphs

The course will be given in English (except if all participants speak French).

• 4:15pm: Simon Dreyer, A Task-Based Taxonomy of Cognitive Biases for Information Visualization -- IEEE TVCG 2018
• 4:35pm: Romain Bouarah, FTK: A Simplicial Spacetime Meshing Framework for Robust and Scalable Feature Tracking -- IEEE TVCG 2021
• 4:55pm: Sacha Cerf Taurus: Towards A Unified Force Representation and Universal Solver for Graph Layout -- TVCG 2022
• 5:15pm: Constantin Gierczak–Galle Crowdsourcing Graphical Perception: Using Mechanical Turk to Assess Visualization Design -- ACM CHI 2010
• 5:35pm: Viviane Ledoux Learning Topology-Preserving Data Representations -- ICLR 2023
• 5:55pm: Eliott Donvez In Pursuit of Error: A Survey of Uncertainty Visualization Evaluation -- IEEE TVCG 2019
• 6:15pm: Victor MIQUEL Vega-lite: A grammar of interactive graphics -- IEEE TVCG 2017

• FibeRed: Fiberwise Dimensionality Reduction of Topologically Complex Data with Vector Bundles -- SoCG 2023
• Visualizing Topological Importance: A Class-Driven Approach -- TopoInVis 2023

• Toward a Quantitative Survey of Dimension Reduction Techniques -- IEEE TVCG 2019

• Visual Analysis of Large Graphs: State-of-the-Art and Future Research Challenges -- CGF 2011
• A Task-Parallel Approach for Localized Topological Data Structures -- IEEE TVCG 2023
• Principal Geodesic Analysis of Merge Trees (and Persistence Diagrams) -- IEEE TVCG 2023
• Discrete Morse Sandwich: Fast Computation of Persistence Diagrams for Scalar Data -- An Algorithm and A Benchmark -- IEEE TVCG 2023
• Branch Decomposition-Independent Edit Distances for Merge Trees -- CGF 2022
• Wasserstein Distances, Geodesics and Barycenters of Merge Trees -- IEEE TVCG 2021
• A Domain-Oblivious Approach for Learning Concise Representations of Filtered Topological Spaces for Clustering -- IEEE TVCG 2021
• A Progressive Approach to Scalar Field Topology -- IEEE TVCG 2021

• DRGraph: An Efficient Graph Layout Algorithm for Large-scale Graphs by Dimensionality Reduction -- IEEE TVCG 2020

TBA