CS 661: Big Data Visual Analytics

Credits: 3-0-0-0 (9)

 

Prerequisites: Basics of Linear Algebra, Statistics, and Probability theory will be beneficial but not mandatory, knowledge in programming (C/C++, Python)

 

Who can take the course: PhD, Masters, 3rd and 4th year UG Students

 

Departments that may be interested: CSE, EE

 

Course Objective:

The necessity of visual analytics capabilities for big data is becoming omnipresent due to its significant demand in the current age of data science and analytics. Interactive data visualization techniques enable us to comprehend and explore diverse types of complex data sets efficiently so that patterns and features from the data can be readily identified and studied in detail. As the data grows larger and becomes intricate, it poses significant challenges to manage, curate, and explore such large data sets in a scalable manner. These data sets can come from various scientific simulations as well as from social media, IoT, various sensors, and many other industry and application domains. In this course, we will cover a comprehensive view of data visualization techniques with a specific focus on the techniques that are applicable to big data. We will discuss the theory and foundations of visualization techniques and have hands-on exercises on visualizing different types of data sets using available visualization software and libraries. We will study scientific and information visualization techniques with a focus on data compression, statistical and information theory techniques, and selected high-performance visualization algorithms. Next, we will discuss how modern machine learning and deep learning techniques are adapted for big data visual analytics. Finally, we will learn about exascale visual computing and state-of-the-art in situ analysis techniques and conclude by discussing the future paradigms of the big data visual analytics domain. The contents for this course will be based on a few books and research papers from top-tier journals and conferences such as IEEE TVCG, CGF, ACM CHI, IEEE/ACM Supercomputing, IEEE Visualization, EuroVis and EuroGraphics, IEEE Pacific Visualization, IEEE LDAV, EGPGV, etc.

 

Course Contents:

 

1. Introduction to visual analytics
2. Data
   1. Different types of data
   2. Big data and its characteristics
3. Foundations of data visualization
   1. Visual perception
   2. Information analysis and visual variables
   3. Data and task abstraction
4. Software
   1. Overview of available visualization software
   2. ParaView, VTK, D3.js
5. Scientific visualization
   1. Scientific data models
   2. Basic visualization techniques
6. Information visualization
   1. Techniques such as Clustering, Dimension reduction, PCP, MDS, SPLOM etc.
   2. High dimensional and graph data visualization
7. Techniques for big data visual analytics
   1. Data compression
   2. Statistical methods
   3. Information theory for big data visualization
   4. High performance algorithms for visualization
8. Machine/Deep learning techniques for big data visualization
9. Data exploration at extreme-scale
   1. Exascale computing
   2. In situ visual analysis
   3. Future paradigm in extreme-scale data visualization

 

Reference:

1. Visualization Analysis and Design by Tamara Munzner, A K Peters Visualization Series, CRC Press.
2. The Visualization Handbook edited by Charles D. Hansen and Chris R. Johnson.
3. Interactive Data Visualization for the Web by Scott Murray.