President's Dream Colloquium on Engaging Big Data

Speaker: Pat Hanrahan

People, Data and Analysis

Pat Hanrahan, Professor at the Stanford Computer Graphics Laboratory
Cofounder and Chief Scientist of Tableau
Founding member of Pixar
Wednesday, February 3, 2016, 7–8:30 pm
Joseph and Rosalie Segal Centre, HC 1400, SFU at Harbour Centre

Recording 

Talk Highlights

00:00 - Welcome by Fred Popowich
02:49 - Opening Remarks by Chancellor Anne Giardini
03:05 - Acknowledgment and introduction to President’s Dream Colloquium
07:10 - Speaker Introduction by Lyn Bartram Pat Hanrahan’s Lecture
09:20 - Introduction
10:02 - Your Typical Data Scientist
11:14 - Application of Data in Jeopardy
12:05 - Convolutional Neural Network, trained in image recognition
14:20 - Predictive algorithm and Hurricane Sandy
16:28 - Data Analysis, Predictive algorithms and serious questions
17:30 - Why we need people in data analysis
19:48 - We need the people in data analysis but what kind?
25:13 - “There are a lot of Data Enthusiasts”
28:05 - Analytical thinking definition
35:11 - Analytical Process
36:17 - First Job
37:18 - Design of Tableau
44:49 - Refining strains for designing Tableau - what to visualize
49:28- The language for visualization
50:17 - Refining strains for designing Tableau
56:44 - The semiology of graphics
58:51 - Principle #1 - Use Semantics
1:00:00 - Principle #2 ­- Good defaults, not rigid rules 
1:01:15 Principle #3 ­- Meta Automation
1:04:20 Wrap Up
1:07:21 Peter Chow-­White Acknowledgment and appreciation talk
1:08:18 Dorcas Yeung’s questions on big data
1:09:32 Dr Hanrahan’s answer to Dorcas Yeung’s questions
1:13:07 Sepideh Hashtroodi’s question on big data
1:13:38 Dr Hanrahan’s answer to Sepideh Hashtroodi’s questions
1:15:19 What’s an analytical assistant
1:17:39 Zahra Zohrevand’s question on big data
1:18:14 Dr Hanrahan’s answer to Zahra Zohrevand’s questions
1:20:12 Audience Q & A
1:36:16 Peter Chow-­White & Dorcas Yeung wrapping up the session

Biography

Pat Hanrahan is the CANON Professor of Computer Science and Electrical Engineering at Stanford University where he teaches computer graphics. His current research involves visualization, image synthesis, virtual worlds, and graphics systems and architectures.  Before joining Stanford he was a faculty member at Princeton.

Pat has also worked at Pixar where he developed volume rendering software and was the chief architect of the RenderMan(TM) Interface - a protocol that allows modeling programs to describe scenes to high quality rendering programs. In addition to PIXAR, he has founded two companies, Tableau and PeakStream, and served on the technical advisory boards of NVIDIA, Exluna, Neoptica, VSee and Procedural.

Professor Hanrahan has received three university teaching awards. He has received two Academy Awards for Science and Technology, the Spirit of America Creativity Award, the SIGGRAPH Computer Graphics Achievement Award, the SIGGRAPH Stephen A. Coons Award, and the IEEE Visualization Career Award. He was recently elected to the National Academy of Engineering and the American Academy of Arts and Sciences.

Lecture

Big data is a hot topic in computing. Most research has focused on automatic methods of data processing such as machine learning and natural language processing. Another important direction of research is how to build systems that can store and process massive data sets. An example here is map-reduce and Hadoop.

Unfortunately, what has been lost in the discussion is how people should use data to perform analysis and make decisions. It is unlikely that people will be replaced completely by automate decision making systems in the near future. Hence, an important question to ask is what should people do and what should computers do? In this talk, I will discuss promising approaches for building interactive tools that allow people to perform data analysis more easily and effectively.

Research Questions

1. When faced with an question involving data, what operations should
computers perform and what operations should people perform?

2. What are some of the most important tasks in data analysis?
How much time and effort is allocated typically to each task?

3. More and more tasks as being automated. What problems can come
up when people rely on automation? How would you design interfaces
that make it easier for people to use smart systems?