"Now, it's like there's almost no area where you can go where there’s not Bayesian statistics:" An interview with Distinguished Professor Kerrie Mengersen

Features

  • Author: Statistics Views
  • Date: 11 Jun 2019
  • Copyright: Image appears courtesy of Professor Mengersen

Kerrie Mengersen studied mathematics and statistics at the University of New England, New South Wales. Her PhD thesis was on the topic of ranking and selection under the supervision of Professor Eve Bofinger, one of the pioneer female university researchers in regional Australia.

Following graduation, she was recruited to a commercial statistical consulting company, which provided her with strong experience in a wide range of statistical methods in the context of diverse applied problems. Her career since then has been characterised by a dual focus of engaging with and developing new statistical methodology motivated by, and motivating, challenging statistical applications.

In 2016, The Queensland University of Technology awarded the title of Distinguished Professor to Professor Kerrie Mengersen in recognition of her outstanding achievements, both nationally and internationally, in mathematics and statistical research. Distinguished Professor Mengersen is acknowledged to be one of the leading researchers in her discipline.

Distinguished Professor Mengersen focuses on using and developing new statistical and computational methods that can help to solve complex problems in the real world. These problems are in the fields of environment, genetics, health and medicine, and industry. She enjoys working with a diverse range of people doing outstanding things in many different areas, and contributing expertise in an important component of their work. Her research interests include complex systems modelling, Bayesian statistical modelling, computational methods and applications, Bayesian networks and applied statistics.

Alison Oliver talks to Distinguished Professor Mengersen about her career so far.

thumbnail image:

1. What was it that first made you interested in statistics and how has this interest evolved during your career?

When I went to university, Statistics was a subject that I just didn't really understand, so I decided to swap to a Maths and Computing degree to learn more about it. The statistics lecturers at the time were a mix of people who taught theory and methods, but then one of the lecturers used to do consulting for different disciplines, such as agriculture, economics and health. He would have students come in from those disciplines and talk about their projects and their problems, and I was able to sit in on that. It included maths, computing, problem-solving and communication. I just thought this was the coolest thing to be able to do.

2. Your research interests include Bayesian statistics, meta-analysis, computational statistics, environmental, genetic and health statistics. What are you working on currently?

One of the great things about statistics is that it cuts through so many areas. I like the combination of the development of the methods and algorithms and then also the application into different areas. In methods, a lot of my work is now focussed on the way that we might analyse big data, but there are also challenges with small data and complicated structures with nonstandard distributions. I am also interested in network models that allow for describing complex systems, new types of data sources and ways of combining different sources of information.

The world is changing in terms of the types of data that we collect. There’s a lot of sensor data, remotely sensed data, citizen science data and so on. It’s fun to invent new ways of using these data sources in statistical modelling. We have projects like building a virtual Great Barrier Reef and allowing divers to be able to geotag their photos to the Reef so that we can access those photos through automatic imaging detection or asking citizens to access and annotate the photos, and we can use that information to improve our statistical models.

We have a similar sort of project where we're using 360 cameras and virtual reality to make virtual reality scenes and putting our experts into those scenes and asking them questions about the probability that a particular rare species might be there and how sure they are about this, and the characteristics of the environment that influence presence. If we have this information for different locations, we can combine this with our observational data to come up with much richer predictions for where these rare species might live. We've done this with rock-wallabies, koalas, and also with creating a Jaguar corridor across the Peruvian Amazon

We also have a project that involves developing an interactive national atlas of cancer for Australia. This will be the first time that Australians will be able to go online and see, for their particular area, the rates of cancer incidence and relative survival, and how these compare to the national average. The atlas is based on Bayesian models that ensure that the estimates are robust and protect privacy.

A lot of potential data sources such as social media and citizen science can contain a lot of noise, but there can also be very useful signals. So, our job is to learn how to identify those signals and tell the story over and above all the noise.

3. You are an elected Fellow of the MSSANZ, IMS and RSS as well as serving as President of the Statistical Society of Australia as well as ISBA. What are your memories when you look back on your Presidency of SSA and ISBA and what do you feel were your main achievements?

I think the professional societies serve a really important role in our professional lives. They’re the way that we can network with others; we can physically come together; they’re a way of disseminating information; therefore they’re, in a way, the glue that binds us as a community of professionals. I've been a member of the Statistical Society of Australia for a long time now, and that's been an important way that I network with the community. The societies are undergoing a lot of change—there are many ways that people communicate now that are not through societies— so part of my role as President was to be able to help raise the profile of the society and to link it with other societies as well.

I think one of the great things is to see young people coming into a society and getting excited about chatting to new colleagues about their own work and creating networks that they will carry through their professional lives. Having an opportunity to shape how we might increase the profile of our community, both for young people and for the existing members and then also to the general public is always valuable. What's the role of statistics, and why might people want to do statistics, and will I get a job afterwards? All of these kinds of questions. I think for the last decade there has been an increasing awareness of statistics in business, industry and government through “big data”.

When I was President of ISBA, one aim was to link to to other international societies so we could raise the profile of Bayesian statistics. Right now, we've been really trying to proactively create safe professional environments and promote respectful professional conduct in the society. We've being developing a code of conduct and procedures to go with that and to proactively address sexual harassment, gender bias, and so on. I think a society has a big role to play in that—it's a way of changing general culture and to also giving everyone the authority to talk about this and an awareness that they can take back to their own institutions. If the society can play a role in supporting them in their own areas of work to really affect this change, then that's really important.

4. You are also the Director of the Bayesian Research and Applications Group (BRAG) as well as the ARC Centre of Excellence for Mathematical and Statistical Frontiers in Big Data, Big Models and New Insights. Please could you tell us more about these roles and the aims of the respective organizations?

BRAG is our research group which that was a late-night acronym that stuck. The emblem is actually a blow-up crocodile called Monty whose teeth look like a type of Monte Carlo traceplot. An iconic Australian crocodile conservationist, Steve Irwin, died in 2008, the same year when Australia hosted the ISBA conference, so we had a blow-up crocodile at the conference. So Monty is ten years old now! It's a great group. We have had a consistently large number of post grads and postdocs all doing different kinds of work in Bayesian stats, and they support each other.

Australia has a Centre of Excellence, funded by our Australian Research Council, in mathematical and statistical frontiers in big data, big models, and new insights. It brings together people from math, stats, and machine learning to work on methods and theory and computation algorithms as well as applications in healthy people and sustainable environments and prosperous society. The aim is to really bring together those three disciplines that were quite disparate but are now becoming more combined, which is exciting to see. The Centre covers seven universities, with funding for seven years, and involves about 100 researchers and about 100 postdocs or post graduate students.

...to be able to inspire people in the application area about the potential that statistics can bring to that discipline, the insights that they can get from their data, but then also to show that this is mathematically sound and that's why it's worthwhile learning. For the maths students, statistics is the way they know of translating that mathematics to applications, and that there are these exciting applications so that they can really have an impact based on their work.

5. In your opinion, what makes a good statistics lecturer? What is the key skill or passion you wish to convey to your students the most?

Passion and expertise. That’s what makes a good teacher. I think being able to tailor the way statistics is taught to the group is vital and very possible—the nice thing about stats is if you are speaking to students that are interested in maths, then it's a very mathematically rich area. If they're more applied, than it can be about the applications and the exciting applications. So, to be able to inspire people in the application area about the potential that statistics can bring to that discipline, the insights that they can get from their data, but then also to show that this is mathematically sound and that's why it's worthwhile learning. For the maths students, statistics is the way they know of translating that mathematics to applications, and that there are these exciting applications so that they can really have an impact based on their work.

It’s really lovely when a student that says, “That’s so interesting. I didn't know you could do that with statistics.” “They thought that stats was this small thing that people did in some back room and they find that there are always really cool things you can do. So, raising that awareness as well as developing the skills makes a good teacher.

6. What do you think have been the most important recent developments in the field?

I think statistics in general has really come to the forefront as a discipline that is such a necessary part of everybody's vocabulary. It’s vital now to have some quantitative skills and awareness of data and data analysis and what data says and how to tell the truth with data. I think having the statistics, machine learning and computer science areas come together and being able to really progress that data science area is extremely important. The other area that’s really been a big push from my perspective—of course, I’m slightly biased—is in Bayesian stats. When I learnt about the area thirty years ago, Bayesian stats years ago was very, very small, and with the increase in computer power, methods, models and packages as well as the awareness of it in what it provides for different disciplines means that it's been taken up at an enormous rate. Now, it's like there's almost no area where you can go where there’s not Bayesian statistics.

7. Your research has been published in many journals and books (over 200 articles): is there a particular article or book that you are most proud of?

I'm proud of articles for different reasons, for their methodological contributions or for their contributions to problems in health, environment or business. I guess two of my earlier ones I'm most proud of, only because I was just learning. When you have your L plates on, I think the first time you drive around the block is always really exciting. Well, it's the same when you first start research: it’s exciting and terrifying. Back then there wasn't a lot of Bayesian stats in Australia, and to be able to publish with people who I thought were amazing in the area was incredible. People like Julian Besag and Richard Tweedie taught me so much about Bayesian stats and helped me really try to understand how these new methods worked. So those papers I still consider to be some of my best.

8. What is the most odd/unexpected topic you’ve written about?

One project that really surprised me was: are there more crimes on full moons? I received a phone call from the police one day when I was at work, and I, “Oh no, what's happened?” This policeman said to me on the phone, “I'm actually ringing because we need some statistics help for one of our people who's doing a project.” And he actually asked whether there are more crimes on full moons. I replied “Really? Are you serious?” Well, it turns out it really is serious, because it influences how police resource different days of the week or times of the year, it affects how emergency services operate and what the hospitals need to provide for, and so on. There is this really big argument in the different communities about whether there is or isn't an effect. So I worked with this guy to analyse 10 years of police call data across Brisbane, where I come from, to see if there was any difference, any real effect, which was incredibly interesting.

9. Congratulations on your election to the Australian Academy of Sciences and the Australian Academy of Social Sciences. You have received several awards and honours and you were also the first woman to be awarded the Statistical Society of Australia's Pitman Medal, which recognises outstanding achievement in the statistics discipline. Is there a particular award or honour that you feel most proud to have received?

I am very proud and still amazed to be elected to the Academies. I'm very proud to be able to represent statistics and all the people that I work with, and to be able to promote science and statistics through these avenues.

The Pitman medal was exciting, for the reason, as you said, the recognition for statistics, but also recognition for women in statistics. We must promote women in science and women in statistics, it's an amazingly good profession for women. For me, statistics combines mathematics and communication and so many different skills within the one discipline, and it ranges across the theory to the application. To able to promote that has been great.

There are many women in all countries that start doing maths and stats at uni. However by the third year the numbers decline and as academics we are very underrepresented. Many leave to work outside academia. One of the things that statistics has for going for it is that there are many different options for where to go after you finish your training, and that is something we can really promote. It can be as flexible as you want, and it can be as competitive as you want.

One award that I received at uni was a trophy that I won for a ‘brass monkey’ squash championship that I won at Walcha, which is a small town in NSW that is extremely cold in the middle of winter. It was a sheer fluke and a funny prize, but I included this award in my CV for quite a while!

10. What would you recommend to young people who want to start a career in statistics?

I would say: go for it. Learn as many quantitative skills as you can and develop them whilst you're at uni or post-uni. It’s never too late to get into this area and I think that there's just such a demand for those skills which will never be wasted. Regardless of what profession you go into, good quantitative skills are such an advantage. In addition to uni courses, there are many online tools and training courses. Some people have asked me, “Should I make a jump into this area?” Yes! And you get some nice emails back going, “I made the jump and it’s great.”

One of the concerns that people have is that their parents don't understand: “Will my child have a job if they go into this area?” Absolutely yes, your child will have a job. They need to do stats, some computing and some coding and if they can get those skills under their belt, then yes, absolutely.

Another very valid concern that I hear is: “What will I do and what are my options?” The answer is everything. There is a wealth of work here in Australia and certainly the opportunity to work overseas. Many, many organisations want data scientists to help them make evidence-based decisions.

To give you some idea of the variety of work, we’re involved with benchmarking organizations who provide services for employment and social service departments, with a conservation group working with orangutans in Indonesia, and with the United Nations on using satellite data.

In the orangutan project, we had a very large survey of 700 villages and 7,000 villagers across Kalimantan in Indonesia. It's hard to know how many orangutans there are and what's causing their demise, so the agency thought: why don't we ask the local people? It was quite contrary to what you might do in ecology at the time—you gather real data, you don't ask people—but we did, and we found out extraordinary answers about reasons for killing, for hunting and fear and all sorts of things and the influence of palm oil and then that's led now to other work on the impact of flooding and palm oil issues in Indonesia.

The United Nations project has involved training official statistics agencies around the world. In addition to traditional surveys, they're looking to use satellite data, wearable data, phone data and scanner data from supermarkets and shops. So even very traditional organizations are looking to see how they can modernize and how they can use these new forms of data, and that is really exciting for everybody to be involved in.

Go and work for a statistics office, you’re not sitting in a dusty room anymore; you can be involved in, “How do I analyse data? How do I track people and understand how people move?” That's important for hospital services and how we build cities and how we devise communities and improve services to rural and remote people and improve health.

11. Who are the people who have been influential in your career?

I’m going to start with my mom and dad. Nobody had gone to university in my family, but they were strong about education and really supported me.

I went to a relatively small rural university, because I grew up in the country, and big cities were a long way away. My supervisor was Eve Bofinger who was one of the few female lecturers in the maths department, and that was a really big influence.


Related Topics

Related Publications

Related Content

Site Footer

Address:

This website is provided by John Wiley & Sons Limited, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ (Company No: 00641132, VAT No: 376766987)

Published features on StatisticsViews.com are checked for statistical accuracy by a panel from the European Network for Business and Industrial Statistics (ENBIS)   to whom Wiley and StatisticsViews.com express their gratitude. This panel are: Ron Kenett, David Steinberg, Shirley Coleman, Irena Ograjenšek, Fabrizio Ruggeri, Rainer Göb, Philippe Castagliola, Xavier Tort-Martorell, Bart De Ketelaere, Antonio Pievatolo, Martina Vandebroek, Lance Mitchell, Gilbert Saporta, Helmut Waldl and Stelios Psarakis.