|
Thursday, February 19 |
SC1 Practical Data Mining: Challenges and Solutions
Fill out evaluation
|
Thu, Feb 19, 8:00 AM - 5:30 PM
Napoleon C3
|
Instructor(s): Richard D. De Veaux, Williams College
Download Handouts
|
|
|
|
|
|
Large data sets (or Big Data) are becoming more common as our ability to collect and store data increases. Many new tools and methods are now available to both the experienced analyst and casual user. Unfortunately, there is a strong belief—due in large part to a series of popular Big Data books—that good results are guaranteed with just powerful algorithms and a lot of data. Instead, success is dependent on the skill and domain knowledge of the analyst and the quality and relevance of the data. However, by using principles of statistical engineering and sound statistical knowledge, the chance of success in these problems is significantly increased. Through a series of case studies, we will show how to be successful in Big Data problems. We will show applications of many current and popular algorithms, as well as when and where they are most successful.
|
Outline & Objectives
Outline:
What is Data Mining and Big Data and where is it being used?
How to pose the problem and get started in Data Mining.
The role of Graphics.
Models -- algorithms and assessment.
Challenges and solutions.
Case Studies -- the role of the statistician as collaborator in data mining.
Objectives: To develop a clear understanding of how to define a data mining problem. Develop the skills needed to attack Big Data problems in a logical and sequential manner, applying principles of statistical engineering and sound statistical thinking.
Clarify the common mistakes of data mining and how to avoid them.
About the Instructor
Dick De Veaux has taught dozens of data mining courses around the world. He is currently the C. Carlisle and Margaret Tippit Professor of Statistics at Williams College where he has taught since 1994. He is a fellow of the ASA, serves on the Board of Directors as a Council of Sections representative and is the newly elected Chair of the Section on Statistical Learning and Data Mining. He is the author (with Paul Velleman and Dave Bock) of Intro Stats, Stats: Data and Models and Stats: Modeling the World and with Norean Sharpe and Paul Velleman of Business Statistics and Business Statistics: A First Course. He is currently working on an introductory book for Scientists and Engineers with Doug Montgomery to be published by WIley.
Relevance to Conference Goals
The course is relevant to all three goals. Clearly it will cover data modeling and analysis and use big data as the main focus. It will also emphasize the importance of teamwork, problem definition and communication.
|
|
|
| |
SC2 From Statistical Consultant to Effective Leader
Fill out evaluation
|
Thu, Feb 19, 8:00 AM - 5:30 PM
Napoleon C1
|
Instructor(s): Roger W. Hoerl, Union College; Ronald D. Snee, Snee Associates, LLC
Download Handouts
|
|
|
|
|
|
This workshop is designed to enhance the leadership skills of statisticians working in business, industry, and government. The goal is to help statisticians transition from being viewed as passive consultants to being viewed as proactive leaders within their organizations. Issues addressed include understanding what statistical leadership is and how it differs from consulting, why it is important to be viewed as leaders, and critical leadership skills required. As part of the course, each participant will develop a personal action plan to enhance their leadership in their own work environment.
|
Outline & Objectives
Workshop Agenda
• Introductions, purpose, agenda and participant expectations
• Why we need to be leaders
• What Management expects of leaders
• What changes are needed by statistics profession and statisticians individually:
o Presentation followed by brainstorming and discussion of brainstormed list
• Effective leadership skills and how to develop them
• Developing action plans for the profession and on a personal basis
• Personalize leadership development plans
• Discussion of selected personal plans
• Wrap up and path forward
Workshop Design
The workshop will use an integration of presentation and discussion of material from critical books and articles on leadership, sharing of personal experiences (participants and workshop leaders) in leading and working with leaders, brainstorming of problems and solutions and development of action plans. The session will be highly interactive enabling extensive participation by all.
About the Instructor
The instructors Ron Snee and Roger Hoerl are well versed in the subject of preparing statisticians to become better leaders. Their publications on Statistical Leadership document their work on the subject. They have presented the proposed workshop on more than 5 occasions over the years.The workshop content, design and delivery have worked well with a variety of statistician audiences including those at the JSM, QPRC and ASA/ASQ Fall technical Conference.
Snee, R. D. and R. W. Hoerl (2004) “Statistical Leadership – as Traditional Workplace Roles Change, Learn to Transition from Consultant to Leader”, Quality Progress, October 2004, 83-85.
Snee, R. D. (2005) “Leading Business Improvement: A New Role for Statisticians and Quality Professionals”, Quality and Reliability Engineering International, Volume 21, 2005, 235-242.
Snee, R. D., R. W. Hoerl and A. N. Patterson (2008) “In with The Right Crowd; Getting Management on Board to Support Statisticians’ Roles and Initiatives”, Quality Progress, May 2008, 70-73.
Snee. R. D. and R.W. Hoerl (2011) “Leadership—Essential for Developing the Discipline of Statistical Engineering”, Quality Engineering, Volume 24, April-June 2012,
Relevance to Conference Goals
This workshop fits nicely with Theme 1: Communication, Impact, and Career Development. The development of leadership and management skills is the purpose of the workshop.Both of the instructors have had leadership and managerial positions in their careers: Snee at DuPont, Bell Atlantic (now Verizon) and Joiner Associates and Hoerl at General Electric
|
|
|
| |