Amy Farris, ASA Director of Membership Development and Marketing, and Donna LaLonde, ASA Director of Strategic Initiatives and Outreach

The American Statistical Association is pleased to announce the Conference for Women in Statistics and Data Science (WSDS), to be held October 20–22 in Charlotte, North Carolina. WSDS 2016 will bring hundreds of statistical practitioners and data scientists together in celebration of women in statistics and data science.

The focus of this conference is to empower women statisticians, biostatisticians, and data scientists by exchanging ideas and presenting technical talks on important, modern, and cutting-edge research; discussing how to establish fruitful multidisciplinary collaborations; and showcasing the accomplishments of successful women professionals.

With leaders from academia, government, and industry, this conference is aimed at encouraging women to enter and stay in these critical fields. The conference environment will be unique and conducive to women sharing and growing their knowledge, influence, and community. Senior, mid-level, and junior stars representing industrial, academic, and government communities will unite to present their life’s work and share their perspectives on the role of women in today’s statistics and data science fields.

The two-and-a-half-day conference will include multiple parallel technical sessions providing participants with the opportunity to learn about novel approaches and innovations addressing the challenges of Big Data. The technical sessions will be complemented by career development sessions for all stages of participants, leadership development sessions, and formal and informal mentoring sessions.

Registration opens June 2, and the housing deadline is September 20. Conference registration ends October 4.

This year’s featured speakers are Cynthia Clark, Stacy Lindborg, Wendy Martinez, and Bin Yu. Here, they reflect on their long careers, share advice for the future, and discuss the topic they plan to talk about at WSDS. Stacy Lindborg was unavailable for the Q&A.

Clark

Cynthia Clark

National Agricultural Statistics Service Administrator (retired)

What or who inspired you to become a statistician?

I did not have a directed goal to study statistics, specifically. I was an undergraduate mathematics major at a liberal arts college for women. I did have an undergraduate calculus-based probability and statistics course. However, my goal was to be a college professor. After earning my bachelor’s degree, I received a Danforth Scholarship for graduate study directed toward teaching mathematics. While in that program, I took a course in econometrics that I really enjoyed. Afterward, I was offered a three-year position as an instructor in the mathematics department at the University of Denver. Realizing I would be limited to teaching calculus unless I had a doctoral degree, I began doctoral studies in mathematics first at the University of Colorado and then—when my husband accepted a position as an associate professor in the Drake University Law School—at Iowa State University in Ames, Iowa. During my first year there as a graduate student, it became clear that there were very few academic positions in mathematics departments, and that a doctoral degree in mathematics might not be a ticket for an academic position.

I realized that the knowledge I had gained from the graduate mathematics courses I had taken was applicable to other fields, so I decided to go shopping for a department at Iowa State where I might pursue my goal of teaching college students. I first went to the computer science department. That was the only time in my career when I felt I was discriminated against as a woman. The department chair was going to make it very difficult for me to be a student in his department. I felt that, as a mother of three preschool children (who was commuting 40 miles from Des Moines to Ames to attend classes), I did not need to face the challenges he was describing. I approached the chair of the statistics department, Professor Bancroft, who was very welcoming. In fact, I could enter without any examination, would have already fulfilled the required mathematics courses, and only needed a statistical methods class, which did not have to be taken before I enrolled in graduate statistics courses. While a student at ISU, my goal continued to be seeking an academic position, now in a statistics department. Only when my husband accepted a position with the Treasury Department in Washington, DC, did I pursue non-academic positions as a statistician.

It was not until much later in my career that I recognized how fortunate I was to have found the highly rated ISU Statistics Department.

Reflecting on your career, what is the most important lesson you’ve learned?

It is hard to pinpoint one lesson. However, in whatever field of endeavor you choose, you need to always continue to learn. My goal has been to improve whatever I am doing or have responsibility for. I set high standards for myself and those who work with me. Much of my career has been as a manager or leader. Early on, I learned there were many work-related problems I did not, myself, have the knowledge or skill to solve. So I had to find others with the desired skills who I trusted to give me knowledgeable advice. As an organizational leader, I often sought advice from multiple individuals who had a diversity of views. I also learned that, as a manager, you always should have more than one person on a project—either working collaboratively or as a back-up in case one of the individuals is no longer able to be on the project. Even senior statisticians should have an individual who reviews their conclusions and work. Succeeding in meeting these challenges requires skill working with people, so that is probably the most important skill to acquire.

Looking to the future, what project are you most excited about?

I recently retired—for the fourth time. I do not plan to commit to another full-time paid or volunteer position. What I am doing is offering my services in an advisory role. I am presently on the boards of the Council of Professional Associations for Federal Statistics, the National Academy of Science’s Panel on Re-engineering the Census Bureau’s Economic Surveys, the Laboratory for Interdisciplinary Statistical Analysis (LISA 2020), and our homeowner’s association. I am a member of the Statistics’ Canada Methodology Advisory Committee and the Washington Statistical Society’s ASA Fellows Committee.

Additionally, I have the opportunity to see my six children and 20 grandchildren more frequently and spend time with them. My husband and I are currently planning a late summer trip with a college-bound grandson to Spain (his choice to use his knowledge of Spanish on the trip). I will continue to pursue my interest in family history, hoping to prepare some biographies of my ancestors. I gained an interest in early Mormon history in a recent 18-month full-time volunteer position in Nauvoo, Illinois, and am currently on a research team with a chaired professor at the University of Virginia adding my knowledge of Hancock County property and geography to her project team.

What advice would you offer an undergraduate statistics major?

One of the things I love about statistics is that I have gained knowledge in many other fields of statistical application as I have studied and applied statistics. You might want to consider a second substantive field as an undergraduate or explore many fields to have a general knowledge of their approach to knowledge and learning. Also, writing and communication skills are very important. Do everything you can to become an excellent technical writer. Working on teams is part of most positions, so honing your skills as a team member is good advice.

What will be the focus of your talk?

I plan to talk about leaving a legacy. What will your legacy be as an individual, a woman, a spouse, a parent, a statistician, a member of society? How do I want to be known or remembered? How do I direct my life now to fulfill that dream?

Martinez

Wendy Martinez

Mathematical Statistics Research Center Director, Bureau of Labor Statistics

What or who inspired you to become a statistician?

It was the early 1990s, and I just completed my master’s degree in aerospace engineering from The George Washington University. I started a new job working as an engineer at the Naval Surface Warfare Center in Dahlgren, Virginia, and my mentor was Carey Priebe (The Johns Hopkins University), who was finishing his PhD in statistics from George Mason University. He told me about a new program they had in computational sciences and informatics, where one of the tracks was in computational statistics. It was an interdisciplinary program that seemed to fit well with my educational background in engineering, mathematics, and physics. Also, statistical methods were at the core of the applications we were working on for the Navy. So, I took the plunge and embarked on a career in statistics.

Reflecting on your career, what is the most important lesson you’ve learned?

I learned it is important to engage in work that one can be enthusiastic about. Therefore, we should not be afraid to try new things in order to remain passionate about our profession. We may not know a lot about a topic at first, but we can learn. So, do not let fear stop you from doing something new and exciting in the vast and ever-changing field of statistics and data science.

Looking to the future, what project are you most excited about?

For the past 10 years or so, I have been interested in the statistical analysis of unstructured text. I started working at the Bureau of Labor Statistics around four years ago, and I found that many offices have many opportunities to use this rich information resource. So, I have been advancing the use of text analysis in surveys and from alternative data sources.

What advice would you offer an undergraduate statistics major?

My advice would be to find a topic in statistics that interests you and learn all you can about it. By their very nature, statistics and data science are interdisciplinary, in my opinion. So, I recommend taking additional classes and electives outside mainstream statistics to expand your knowledge. For example, having a background in computer science, artificial intelligence, computational linguistics, mathematical modeling, and/or data mining would provide a useful skill set for a new statistician.

What will be the focus of your talk?

My talk will focus on Big Data in the federal government and some related grand challenges faced by statistical agencies. I plan on talking about some of my experiences with Big Data over the past 10 years, starting with my time at the Office of Naval Research and ending with my current position at the Bureau of Labor Statistics. I will provide several research challenges to help motivate statisticians and data scientists working in this exciting area.

Yu

Bin Yu

Department of Statistics and Department of Electrical Engineering & Computer Science Chancellor’s Professor, University of California at Berkeley

You have a joint appointment in the departments of statistics and electrical engineering and computer science. How did you become involved in interdisciplinary work?

First of all, a big part of interdisciplinary research is to work with many other people, a number of whom are scientists. Getting to know them and learning science from them are the most exciting and rewarding parts of interdisciplinary research for me, especially when there is good synergy at both scientific and personal levels. To honor their contributions, I mention below names of all my main collaborators, including students and postdocs in the past years.

When I was a PhD student at Berkeley in the late ’80s working with Terry Speed and Lucien Le Cam, as a pioneer of statistical bioinformatics, Terry was beginning to collaborate with biologists on biological problems in a serious way. Many of my friends were Terry’s students working on bioinformatics, as well. They were sitting in an upper-division basic genetics course and I sat in with them. I was going to all the talks on bioinformatics and picking up stuff along the way. In the summer of my fourth year, I worked on lipoprotein data with Ron Krauss at Lawrence Berkeley National Laboratory under Terry’s supervision. We used EM algorithm to find different sub-populations of patients with different HDL and LHL profiles, and it was my first interdisciplinary project, although my thesis work was theoretical and on empirical processes and information theory. The part of the thesis on information theory was actually interdisciplinary already, but theoretical. After my PhD, I was introduced to the information theory community by Jorma Rissanene, an IBM fellow and inventor of MDL (Minimum Description Length Principle). Jorma was my third PhD adviser in some sense, since he came to Berkeley every month and worked with me and Terry. I was awed by both the beauty and usefulness of Shannon’s Information Theory. I became an active member of the information theory community and was welcomed there. I got to know other information theoreticians such as Tom Cover, Jacob Ziv, Imre Csiszar, and Sergio Verdu. This information theory theoretical interdisciplinary connection prepared me to get into signal processing later.

That summer work with Ron happened because I had an interest in applied work, since I reasoned that most of the good and creative ideas in statistics seemed to have come from solving real problems at the boundary of statistics and other fields. Before the summer, I had been reading Fisher with Terry and taking his applied statistics classes. He was wonderful as a mentor or adviser—going to the library with me, inviting me to lunch at his house every Saturday, and telling me about all his applied statistics projects—and answering my questions about them. I picked up from him a lot of “data wisdom”—a term I have coined for the essential elements of applied statistics in a web article at a Big Data website, obdms.org. Then Terry provided the opportunity with Ron in that summer. I was paid by Ron as an RA. Many years later, Ron got in touch to find statistical expertise for his current job at Oakland Children’s Hospital and I introduced my colleague Haiyan Huang to him. They are still working together.

After my PhD, I went to Wisconsin-Madison as an assistant professor. One reason for this choice was to be influenced by George Box or the empirical style of English statistics. Unfortunately, I did not get to interact much with George since he was retired when I got there. I did look for opportunities to do interdisciplinary, but nothing panned out.

I returned to Berkeley in the fall of 1993. Around 1995, I attended a Neyman seminar by Martin Vetterli on wavelet signal processing. I thought it was really cool and talked to him after the seminar. He graciously invited me to attend his weekly group meeting and introduced me to his former student, Antonio Ortega, with whom I wrote my paper on wavelet image compression. My first student, Grace Chang, was joint with Martin.

“Follow your passion and learn how to learn on your own, since you need skills to realize your passion. You cannot learn all the skills in college that are needed in the future, since science and technology move very fast.”

In later years, I was at Bell Labs from 1998–2000 and worked on network tomography with colleagues Jin Cao and Scott Vander Wiel and low-delay and low-complexity speech compression with Gerald Schuller and Dawei Huang. Then, I engaged in remote sensing research with colleagues Amy Braverman, Eugene Clothiaux, Ming Jiang, my student Tao Shi, my joint student Xin Jiang with Ming, and postdoc Ethan Anderes for cloud detection at the polar regions. For the aerosol retrieval project based on multi-angle satellite (MISR) images, I worked with Yang Liu, my Berkeley student Nancy Wang, and postdoc Taesup Moon.

My more recent interdisciplinary experience is described below after the next question.

Reflecting on your career, what is the most important lesson you’ve learned?

Hold oneself up to one’s own values and standards.

Looking to the future, what project are you most excited about?

That is a hard question, since I have at least four projects I am very excited about. They are also very different, so I can’t order them—that won’t do justice to them. If I may, I would like to say something about all of them.

First is a long-term collaboration with Berkeley neuroscientist Jack Gallant’s lab on understanding a challenging visual cortex area V4 using deep learning or convolutional neural network (CNN) with my students Reza Abbasi, Yuansi Chen, and Adam Bloniarz. We are writing a paper called “Artificial Neurons Meet Real Neurons: Pattern Selectivity of V4.”

The second is also a long-term collaboration with biologists Erwin Frise and Sue Celniker of Lawrence Berkeley National Lab (LBNL) that uses novel spatial gene expression data to understand how organs are formed in the modern organism Drosophila with my students Siqi Wu and Karl Kumbier and former postdocs Antony Joseph and Siva Balakrishnan. We also work with Wei Xu’s computer science team at Tshinghua University to scale up the computations by building upon open-source platforms Spark and Fiji. This is my favorite data science project since it represents an iterative knowledge discovery process that is complete with wet-lab knockout experiments, statistical and machine learning methodology development, and software development for other groups to go after heterogeneous building blocks hidden in their data, spatial or not. This project also motivated exciting theoretical work on dictionary learning. The theoretical study has made us go back to practice for the next step of devising uncertainty measures. It would not have been possible without my amazing student, Siqi Wu.

The third is a collaboration with computational biologist Ben Brown of LBNL to discover nonlinear interactions between biomolecules using iterative Random Forests (iRF). We are writing a paper with our joint postdoc Sumanta Basu. This project has motivated new theoretical nonlinear regression models that we put into a proposal.

The last is a beginning project with my Berkeley colleagues Jas Sekhon and Peter Bickel on heterogeneous effect estimation in causal inference and precision medicine. This project is powered by graduate students, Soeren Kuenzel, and Rebecca Barter.

We are also using random forests here, so good synergy with the third project at the methodological level. By the way, all the projects use state-of-the-art nonlinear methods CNN or RF or dictionary learning, which are at the frontier of statistics and machine learning as causal inference’s mixing with machine learning.

What advice would you offer an undergraduate statistics major?

Follow your passion and learn how to learn on your own, since you need skills to realize your passion.

You cannot learn all the skills in college that are needed in the future, since science and technology move very fast.

What will be the focus of your talk?

I have not made the final plan yet, but I think I will speak about the fruit fly project, discuss my understanding of how good research comes about, and share lessons learned as a woman scientist engaged in interdisciplinary and theoretical research for decades.