Newsletter Spring 2013


Newsletter Stories:

 

Statistics Alumna Creates New Scholarship for Undergraduate Students

Statistics alumna Leslie Cavarra Buttorff has donated a $100,000 cash gift to the Department of Statistics at Colorado State University to help students achieve degrees in statistics, which is a key component to Buttorff’s personal success. “I have used Statistics successfully in my career to think differently about things, and to solve problems,” said Buttorff. “Probability and analysis come into play in almost every business opportunity we are involved with.”

The Department is in the process of reinstating the undergraduate Major in Statistics, a goal Buttorff strongly supports. The scholarship is intended to allow an undergraduate student majoring in Statistics to receive $20,000 per year. The aim is to help a student stay in their Statistics degree program. “I hope the scholarship will enable someone like me to obtain a great and enjoyable education in the Statistics Department,” Buttorff said. “When I was in high school, we did not have a lot of money and I could not really afford to go to college,” said Buttorff. “Hopefully, this scholarship will allow someone to go to college who would struggle otherwise.”

Buttorff is CEO and President of Quintel Management Consultants, which is a company she founded in 2002 that deals with management, consulting and technology integration. Quintel is a woman-owned business focusing on performance improvement and SAP implementations. “From my first job out of college until now, I have used applied statistics frequently in my career,” said Buttorff. “I spent 15 years at Stone and Webster working on regulatory matters for many utilities across the globe. Statistics was used to determine increment pricing, marginal power demands, and failure rates for nuclear and hydroelectric equipment. Statistical distributions are used for many applications in maintenance and procurement planning.”

Her degree in statistics helped her become successful in the business world. She received her bachelor’s degree from Colorado State and her master’s degree from Iowa State University. She hopes to see more statisticians come from CSU in the future. “I think statistics allows one to combine business, math and economics all into one subject and this is valuable for the business world,” said Buttorff. “Statistics teaches you to solve problems. When we hire Stat majors, I know we will get a person that is probably more innovative and less structured than a math major.”

Math had always been Buttorff’s favorite subject in school. She knew that she wanted to apply math to her future career somehow. “I first remember being interested in math when my first grade teacher had contests using flash cards. I remember I liked to win,” said Buttorff. It wasn’t until high school, however, that Buttorff was introduced to statistics. While attending Wheat Ridge High School in the Denver metro area, a CSU representative came to Buttorff’s math class and talked about statistics degrees. She decided to major in statistics at CSU.

“I really did not know much about the degree until I started taking classes. I just knew I wanted to apply my math aptitude,” said Buttorff. While at school, Buttorff was influenced by her advisor, Tom Boardman. “He helped me get work in the Stat Lab. The work that we did in the lab was very interesting and made me even more excited about staying with the degree,” said Buttorff.

For a longer version of this story, see the CSU press release here.

 

 

Message from the Chair, Jean Opsomer

Dear Alumni and Friends of the Department of Statistics,

Welcome to this second edition of our revamped newsletter.  We plan to continue publishing our newsletter in this new format twice a year, at the end of the fall semester and in early summer.  We had a busy six months since the last newsletter, and I encourage you to read some of the stories in this newsletter.  But first, let me talk briefly about some recent department happenings.

By the time you read this, we will be holding the Graybill 2013: Modern Survey Statistics conference in Fort Collins.  With a few minor gaps, this conference series has been on-going since 2001, and is a great opportunity to welcome world-class researchers, students and interested professionals to beautiful Fort Collins in June.  Jay Breidt and myself co-organized the conference this year, on a topic that is a particular research interest for both of us.  We are delighted to have been able to put together a program that features many of the pre-eminent survey statistics researchers in the world.  The next two Graybill conferences are already in the works as well, so stay tuned for future announcements on those.

The last two years, the Department has hosted a well-attended “CSU Department of Statistics Alumni and Friends” reception at the Joint Statistical Meetings.  Next August, we are again planning a reception at the JSM, so mark your calendars: Tuesday August 6, 5:30-7:30pm, Intercontinental Montréal, room Saint-Alexandre.  Reconnect with past and current faculty, staff and students of the Department, pick up some CSU swag, meet new friends!  An invitation with RSVP will follow in the next few weeks.

We already mentioned the efforts to reinstate the undergraduate Major in Statistics in the previous newsletter.  So while not news exactly, I am pleased to report that our diligent pursuit is starting to bear fruit.  The reinstatement was tentatively approved by the CSU Council of Deans last month, and we are now working on the final proposal, which will eventually have to be approved by the Colorado Commission on Higher Education.  We are hopeful that the Major will be available again to undergraduate students in spring 2014.  Many of our alumni have felt strongly that the discontinuation of the undergraduate Major was a loss for CSU, so we are delighted that we are now able to reverse that decision.  As the above story about alumna Leslie Buttorff clearly demonstrates, the CSU Statistics degree is a great stepping-stone for a successful career in a quantitative field!

Let me close by commenting on the prospects for the University and the Department.  Following several difficult years during which decreases in state support resulted in substantial budget cuts throughout the university and hefty tuition increases for students attending CSU, the institution is now on a much more solid financial footing.  In order to remain financially viable into the future and compete in the global higher education marketplace, President Tony Frank is embarking CSU on an ambitious growth path, projecting to increase the CSU student population by approximately 50% over the next 10 years, bringing us to 40,000 on-campus students.  In order to accommodate this growth, CSU will need more faculty, more staff, more classrooms, more student living areas, more research space, etc.  As this expansion reaches the various CSU colleges and units, we expect to grow the Department of Statistics as well, which gives us the opportunity to broaden the range of courses we offer, increase the number of graduate students we can support, and participate in new interdisciplinary programs on campus.  In short, it’s a great time to be at CSU!

I look forward to seeing you in Montréal if you plan to attend the JSM.  If not, don’t hesitate to contact me, I very much like to hear from our alumni and friends.

Jean

 

 

Department Finished First Year of New Master of Applied Statistics (MAS) Program

Only a few months remain of the inaugural MAS year, which has been very successful.  A total of 44 students are enrolled in the program, and seven on-campus students will complete the course at the end of June and be our first graduating class.  The 37 distance students are enrolled in the program on a part-time basis and will typically take between two and four years to complete the degree.  We expect to see further growth in the program next year.

The program started in August with an intensive “bootcamp” to review concepts of calculus and linear algebra and to introduce the students to statistical computing with SAS and R.  Now, less than 10 months but many hours of instruction and hard work later, the on-campus students are completing courses on topics such as longitudinal data analysis, generalized linear mixed models and Bayesian hierarchical models.  This is an impressive accomplishment, and both the instructors and the MAS students deserve ample credit for it.

Starting in May, the MAS program is culminating with the capstone consulting module.  This six-week course provides our students with exposure to a real-world consulting experience and presents an opportunity for the application and consolidation of methodologies acquired over the year.  The projects range from industrial applications with Colorado and Wyoming companies, to analysis of health care and political data for researchers at CSU and local organizations.

On reflection, from an instructor’s perspective, the most difficult aspect of the MAS program has been the fairly rapid pace of instruction, with 19 courses in less than one year, and the need to ensure that all the students are able to keep pace with the materials.

At the end of the spring semester, all faculty involved with the MAS program met to review the first iteration of the course. Crucially, and in addition, detailed feedback will be sought from those students completing the degree.  We look forward to welcoming our second on-campus class next August and continuing to work with our distance students.

 

 

Graduate Students Complete Their Degrees, Look Forward to New Challenges

The year sees a true “bumper crop” of MS and PhD students finishing their studies, with six MS and five PhD students graduating.  While we are of course sad to see them go, we are also delighted that we were able to train so many new statisticians, who will shortly be going to a broad range of industry, academic and government positions all over the US.  This is not the place for an exhaustive list of where everyone is going, but let us mention some of them below.  

Starting with the MS students who will be staying the closest, Ben Prytherch (MS) is starting as a full-time instructor in our Department.  He will be teaching several new sections of STAT301, one of our popular introductory statistics courses that have suffered from long waiting lists in recent years.  Next is Barb Andre, current CSU employee in the Department of Microbiology, Immunology and Pathology, who pursued her MS part-time and will transition to full-time Statistical Analyst in the same department.  Finally, Alison Cartwright is also completing our MS degree and staying at CSU, pursuing her PhD in the interdisciplinary Graduate Degree Program in Ecology.  Going somewhat further afield (but just barely!), David Fox and Kristin Adachi are both finishing the MS degrees and taking positions as statistical consultants with West, Inc. in Cheyenne, WY.

Among the PhD students, Huan Wang is the only one going to industry this year: she is joining Eli Lilly Global Statistical Sciences group in Indianapolis, IN, supporting diabetes research. She was co-advised by Mary Meyer and Jean Opsomer.  Erin Schliep was advised by Jennifer Hoeting and is starting as a postdoc at Duke University, with Alan Gelfand and Jim Clark.  Ephraim Hanks (advisor: Mevin Hooten) starts as assistant professor at Pennsylvania State University, while Grant Weller (advisor: Dan Cooley) will be assistant professor at Carnegie Mellon University.

We wish them all the very best in their new endeavors!

 

 

Undergraduate Student Profile: Melissa Francis

Everyone in the Department knows Melissa Francis, because while pursuing her studies, she also worked in the main office of the Department of Statistics for the last three years. Melissa is completing two degrees in four years with three majors: statistics, general mathematics, and economics.  While working towards these three majors, she put herself through school working a minimum of 35 hours per week.  In June, she starts her new job at Kaiser Permanente in Denver, where she will join the IT/software engineering group.

While pursuing her degree, Melissa definitely made the most of her time at CSU.   Melissa served as a CSU Ambassador for the College of Natural Sciences and the Mathematics representative for the CNS College Council.  She also got involved in many different research experiences.  Melissa worked with Jean Opsomer doing data analysis for a National Marine Fisheries Service project.  The focus of this work was to investigate alternative sample allocation scenarios for recreational angler surveys conducted by the California Department of Fish and Wildlife.  Melissa did the R programming for the research.  Melissa also helped Jean write an article that was featured on the cover the April 2013 issues of Amstat News. In that article, Melissa and Jean analyzed the presenter satisfaction surveys conducted after the Joint Statistical Meetings in 2010-2012.

Melissa and Ben Prytherch, an MS student in our department, worked this spring on a project in our co-taught graduate and undergraduate course on Statistical Consulting (STAT586/STAT472).  In this course, graduate students mentor undergraduate students on statistical consulting projects with real clients.  The students analyze complex data sets to solve real-world problems.  Melissa and Ben worked on a project for the Fort Collins Police Department, which investigated whether there are any significant persistent patterns in crime over time in the Fort Collins area (the punchline: not really). 

Melissa’s interest in programming, which eventually led to her job at Kaiser, was sparked by her exposure to the multiple programming opportunities in her statistics major.  Melissa used R Statistical Software in Stat340 (Multiple Regression Analysis) STAT372 (Data Analysis Tools) and STAT472 (Statistical Consulting).  She also was the main programmer while doing research in FEScUE (http://www.fescue.colostate.edu/), an interdisciplinary undergraduate research experience for Statistics and Mathematics majors.

 

 

Jay Breidt Named Professor Laureate of the College of Natural Sciences

For their outstanding contributions in research, teaching and outreach in their respective fields, Jay Breidt, together with Pat Bedinger (Professor, Department of Biology) have been named the 2013 Professor Laureates by the College of Natural Sciences.  Professor Laureate is the highest academic title awarded by the College of Natural Sciences to select faculty for their career accomplishments.  “Professors Bedinger and Breidt epitomize the scholarly excellence that impacts students in a significant way while contributing to the community,” said Jan Nerger, Dean of the College of Natural Sciences. “They join the highest ranks of other Professor Laureates who demonstrate the highest level of research, teaching, mentoring and outreach.”

Breidt’s research interests include time series and survey sampling.  In addition to researching new theory and methods for nonparametric regression estimators in surveys, supported by the National Science Foundation and the US Forest Service, he has helped revise recreational fisheries surveys for the National Oceanic and Atmospheric Administration, advised South Africa on the development of a new land resource inventory, and assisted the state of Colorado in election auditing procedures. Recently, Breidt has been a member of the Federal Economic Statistics Advisory Committee, which guides analysis and development of national statistics. His work helps coordinate efforts of three major agencies: The U.S. Census Bureau, the Bureau of Labor Statistics and the Bureau of Economic Analysis. He works to examine the agencies' programs and provides advice on statistical methodology, research needed and other technical matters related to the collection, tabulation and analysis of federal economic statistics.

As we highlighted in our Fall Newsletter, Breidt is also currently supported by the National Science Foundation and the National Institutes of Health as the PI of a multidisciplinary study aimed at identifying the shapes of complex biological molecules from small-angle X-ray scattering (SAXS) experiments. This problem is fundamentally important in discovering how the molecule functions with respect to human health and disease.

The title, Professor Laureate, provides recipients an honorarium and three years of resources for student projects.  Each Professor Laureate delivers a public lecture on his or her research. Breidt’s lecture is scheduled for the fall semester of 2013.

Geof Givens Helps International Whaling Commission, Inupiat People Determine Quotes

The Inupiat people native to northern coastal Alaska – one of the most isolated spots on the planet – have hunted bowheads near Barrow as part of their culture for millennia. That hunting has been regulated in recent decades by the International Whaling Commission, which controls who hunts whales – and where – around the globe.

As part of his research, Geof is estimating the abundance of the western Arctic bowhead whale population for these villagers. He will have a preliminary answer in June.  No ice-based count of these whales has succeeded since 2001.  Scientists perch on the sea ice edge in harsh conditions continuously for two months to count the whales and track them by recording whale song. However, more than three quarters of whales swim past undetected, Geof suspects.

“We’ve got unprecedented data from sea ice observation posts and underwater microphones so now it’s up to me and graduate student Stacy Edmondson to estimate total population size,” said Geof. “It’s tricky because these whales are often swimming under the water or ice. We need to estimate how many whales we never saw or heard, then add the number that were detected.”

After Geof’s team presents a final estimate to the International Whaling Commission in June 2014, the future of the Inupiat hunt depends on politics.  Once the International Whaling Commission issues an official ruling on the quota, the decision is legally binding. Each country that participates in the commission has one vote.  Geof’ research has earned him a spot as a US delegate to the scientific committee of the IWC for two decades.  More than 80 countries belong to the IWC. Only about three statisticians serve in the scientific committee.

Typically, the Inupiat people have been allowed to harvest about 55 whales per year out of a population of about 13,000 whales in the region – a fraction of 1 percent. That population is growing at about 3.5 percent a year, “making it a win-win for both conservationists and hunters,” Geof said. Alaska also shares five whales in its quota with Russia.

In 1848, Yankee whalers discovered these whales and almost annihilated them, which devastated a resource that has been a critical food source for the coastal Inupiat and a pillar of their culture.  In the 1970s, the IWC became concerned about bowhead conservation and the North Slope Borough began a comprehensive research program, culminating with Geof’s project. The North Slope Borough governs the vast region of northern Alaska inhabited by the Inupiat.

The North Slope Borough research program and Geof’s work are funded from the Borough budget, the Alaska Eskimo Whaling Commission, the National Oceanic and Atmospheric Administration, and oil company BP.  It is unique in how native hunters, science and industry share information and resources.

CSU’s online magazine “Today” has a longer version of this article, including photos of whales and scientists in this arctic region; see http://www.today.colostate.edu/story.aspx?id=8453

 

 

2013 is International Year of Statistics

The International Year of Statistics, also called Statistics2013, is a year-long celebration of the discipline of statistics, intended to raise the awareness of this profession and the ways in which it provides advancements to our society.  Our department and nearly 1900 other organizations from 122 countries are taking part in this event.  In addition to raising awareness of the discipline of statistics, other objectives of this campaign are to encourage individuals, particularly young people, to consider statistics as a profession and to promote creativity and development within probability and statistics.  Our department is participating by linking two public events to Statistics2013 – International Summer Scholar Presentation "Statistical Network Science" by Ernst Wit on May 23, 2013 and the 2013 Graybill Conference: Modern Survey Statistics, June 9-12, 2013.

You can learn more about the International Year of Statistics at www.statistics2013.org, a website which includes information about statistics, career information, teaching resources, statistics in the news from around the world and many other resources.

 

 

Department Introduces New Course to Prepare PhD Students

Don Estep, Professor of Statistics and Mathematics, is teaching a new course that aims to bridge the gap between the Master’s level and PhD level theory curriculum.  The course is currently an experimental course and as such, it has a rather obscure designation in the university catalog.  But everyone in the department affectionately calls it “STAT620,” since it sits between the MS-level STAT520 and the PhD-level STAT720.

The idea behind STAT620 to build a bridge between the strongly intuitional material in the introductory probability course STAT520 and the heavily theoretical material in abstract probability in STAT720. The possibility of generalizing from particular examples to an abstract result that applies to a wide range of problems is one of the key features of the mathematical sciences.  “Yet in my experience”, said Don, “graduate courses drive to the abstract level too quickly, and students suffer because they are not given time and the ingredients to develop an intuitional grasp that is needed to take advantage of abstract understanding. In my experience in teaching analysis, students who learn analysis at the gut level stand the best chance of using the ideas in their own work later on.”

This situation strikes Don as ironic because most important mathematical ideas are built on strongly practical grounds. For example, the heart of probability and measure theory is based on approximations of complicated sets and functions by simple sets and functions and the idea of describing probability in terms of ratios of areas. The simplest key results and ideas in measure theory probability can be explained in concrete terms that make “computational sense,” and are actually quite fun to learn.

STAT620 is aimed at giving our graduate students a strong intuitional grasp of probability and some key ideas of analysis.  “I have enjoyed teaching the ‘first draft’ this semester,” says Don. “The material is fun and I can spend a lot of time explaining ideas, showing examples, asking the students questions. I grade the course based on take home problem sets, where students are allowed to redo questions that they miss. So I think the students get a lot from the course. I will not get as far as I planned, mostly because I had to do quite a bit of early exploring to get a sense of what the typical student remembers from undergraduate analysis. Next year, I will be able to a better job defining the syllabus in order to present a more relevant range of materials.”

 

 

Springer Publishes a Monograph on Functional Data by Kokoszka and Horvath

Last year, Piotr Kokoszka, Professor of Statistics, published a monograph on functional data with Lajos Horvath, Professor in the Department of Mathematics at the University of Utah. The book is entitled “Inference for Functional Data with Applications,” and has been published in the prestigious Springer Series in Statistics. The book is about 450 pages long and consists of 18 chapters covering recent developments in the field of functional data analysis, with emphasis on dependent data and methods based on functional principal components.

In functional data analysis, complex objects like curves or surfaces are viewed as individual, indivisible observations rather than as collections of numbers that make them up. For example, an instrument may record data every five seconds, and a researcher may be interested in how the shapes of daily records evolve from day to day.  Data of this type abound in many applications: daily price or yield curves, annual temperature or precipitation curves, brain scans (images rather than curves) of many subjects or of a single subject over time, etc.  As the sheer amount and complexity of available data continue to increase, functional data analysis methods are an important and fairly new approach to address the inevitable challenges of how to properly model them.

According to Piotr, the book can be read at two levels. Readers interested primarily in methodology will find detailed descriptions of algorithms, together with their assessment by means of simulations studies, followed by examples of application to real data. Researchers interested in mathematical foundations of these procedures will find carefully developed asymptotic theory. The book provides a concise introduction to the requisite Hilbert space theory, which many graduate students or advanced undergraduate students may find useful.

Piotr joined the Department of Statistics at CSU in August 2011. Previously, he was a faculty member at Utah State University.  He has published over 100 papers in various areas of statistics, including asymptotic theory, functional data analysis, time series analysis, wavelet and Fourier analysis, applications to econometrics and geosciences, mostly space physics.   He serves on the editorial boards of three journals: Computational Statistics, Journal of Time Series Analysis, and Statistical Modeling.

Citation: Horváth, L., & Kokoszka, P. (2012). Inference for functional data with applications (Vol. 200). Springer.

ISBN: 978-1-4614-3654-6

 

 

Don Estep is Founding Co-Editor of New ASA-SIAM Journal on Uncertainty Quantification

American Statistical Association (ASA) and Society for Industrial and Applied Mathematics (SIAM) jointly agreed to start the new “SIAM/ASA Journal on Uncertainty Quantification” (JUQ for short) in January 2012. It is the brainchild of Don Estep, Professor of Statistics and Mathematics, and Jim Berger, Professor in the Department of Statistical Science at Duke University, who serve as co-Editors in Chief.  The Journal began receiving submissions in May 2012, and the first edition appeared in March 2013.

Many statisticians react with puzzlement when they hear there is a new journal devoted to “uncertainty quantification (UQ)” – after all, quantifying uncertainty seems to the main goal of statistics. But, UQ encompasses a much wider range of aspects than the statistical components.  Don became interested in UQ from the mathematical side, and according to Don, his growing involvement in this area was one of the main reasons he moved to statistics a few years ago.

Don explains this transition in his research: “For many years, I worked on computational error estimation for numerical solution of deterministic models. The goal of the work was often to provide information to a user about the reliability of model results. The nature of modern science and engineering, which relies increasingly on model simulations, made quantifying accuracy of results paramount in a wide range of applications.  But in my many interdisciplinary interactions, I realized that while numerical error in real applications is always significant, so too was error and uncertainty in data and parameter values obtained from experiment. I had no tools to deal with that directly. As I began to investigate the aspects of error arising from data, I encountered the statistical side of the problem. Here, many statistics-based researchers had tools for dealing with data error and uncertainty, but no systematic way to deal with deterministic discretization and model errors.  So I started employing both statistical and mathematical approaches to deal with the general UQ problem.”

The problem, however, was that it was often difficult to find good places in the existing journals in which to publish. The goals and analysis, even the very orientation of the problems, was far outside what is usually considered in the disciplinary journals.  At the same time, many people in both statistics and mathematics were beginning to study related problems and questions and finding a natural home difficult.

Don describes the origins of JUQ: “The seeds of the JUQ were laid down in discussions with Jim (Berger) and Max (Gunzburger) over drinks at various conferences featuring sessions on UQ where we met. The momentum built up during a National Town Hall Meeting on UQ we organized at SAMSI for SIAM and ASA.  Starting a new journal in the professional societies ASA and SIAM is a major event, and the proposal process took well over a year. There were many evolutions of the early ideas required to adapt to objections and suggestions raised by various review levels in both societies. But, in the end, the proposal was approved with great enthusiasm on both sides.”

For further information, here is a link to SIAM press release about first volume: http://connect.siam.org/siamasa-journal-on-uncertainty-quantification-launches/ and a link to journal website: http://www.siam.org/journals/juq.php.