American biostatistics professor Liz Stuart, who was born in the Sydney suburb of Kogarah, drew an audience of about 60 people on 21st October 2019. After running a full-day and fully-subscribed short-course earlier that day she managed to stay on her feet all the way to the end.

An early example in Liz's talk concerned a health insurer that may be deciding to approve a new treatment for back pain. Of interest is the "population" average of the treatment effect.

Some of the themes of the presentation were causal inference, generalisability. For example, results from randomised trials may not generalise to the population of interest. Liz then projected a rare technical slide, concerning her work on population average treatment effect.

Problems such as biases if participation in a trial is associated with impacts were mentioned. For example, large urban school districts in the United States are more likely to participate in education studies.

Liz said that often one has to deal with the reality of only having a single randomised trial. Remedies based on Bayesian additive regression trees were discussed.

Towards the end of the presentation an example concerning highly active retroviral therapy for patients with human immunodeficiency virus was discussed. A problem was the high proportion of people in the trial being older, white and male and this tended to bias the population average treatment effect. The problem of generalising beyond the study was brought up again.

A closing quote, which summed up much of what Liz spoke about was "You can't fix by analysis what you bungled by design".

Matt Wand
University of Technology Sydney

SSA is pleased to announce the inaugural Betty Allan Travel Award. Named after CSIRO’s first statistician, this award ($3000) will support travel of one early career female Stats Society member and/or CSIRO/Data61 staff member currently working or studying in the field of statistics in Australia, to a suitable location and/or conference anywhere in the world. This award will be available once a year.

It is jointly funded by the Statistical Society and CSIRO. Applications are now open and will close on Friday, 29 November 2019.

For more information about Betty Allan and this award, please go to the SSA website.

From: Medical Journal of Australia (MJA)

A LACK of attention to biostatistics as a core scientific discipline threatens the value of the $800 million spent annually on Australian health research investment, in terms of improved health and lives saved, according to the authors of a Perspective published online today by the Medical Journal of Australia.

“The entire Australian medical research enterprise is at considerable risk of ‘drowning in data but starving for knowledge’,” wrote the authors, led by Associate Professor Katherine Lee and Professor John Carlin, of the Clinical Epidemiology and Biostatistics Unit at the Murdoch Children’s Research Institute.

Biostatistics “provides the theoretical basis for extracting knowledge from data in the presence of variability and uncertainty”, they wrote.

“It is a critical element of most empirical research in public health and clinical medicine, with the best studies incorporating biostatistical input on aspects from study design to data analysis and reporting. Biostatistical methods underpin key public health research disciplines, such as epidemiology and health services research, a role that reflects the core nature of the discipline of biostatistics.

“Superficial understanding of statistics can easily lead to unscientific practice (recently characterised as ‘cargo-cult statistics’) and may be seen as responsible in large part for the current ‘crisis of reproducibility’ in research.

"The emerging era of big data heightens the need for biostatistical expertise, with more decision makers and researchers aiming to extract value from complex messy data."

In the US, the UK and continental Europe major universities have established departments of biostatistics, or have national centres in biostatistical methodology, as well as dedicated streams of funding for methodological research, they wrote.

In Australia, however, “there has never been systematic investment in the development of biostatistics … either in universities or via national funding schemes”, Carlin and colleagues wrote.

“None of the major universities has a department of biostatistics.”

The authors suggested three potential solutions:

•       universities and research institutes need to foster the development of organisational structures with a critical mass of academic biostatisticians working both in methodology and collaborating with health researchers, as well as training opportunities and career development for biostatisticians;

•       biostatistical teaching and advanced training must keep pace with the dramatic changes in the data science landscape, to ensure that graduates have the necessary breadth of skills to support medical research in the modern era; and

•       funding bodies need to invest in biostatistical research; for example, by the creation and support of graduate and postdoctoral methodological training programs.

Please remember to credit The MJA. The Medical Journal of Australia is a publication of the Australian Medical Association.

On 24th September 2019 the New South Wales branch gathered on the

campus Macquarie University in north-western Sydney. Members of

the council with offices in the inner suburbs had their Opal cards

and transport apps kept busy in the race to out to North Ryde for

their pre-talk monthly meeting. The writer of this article was declared

the winner, making it door to door in just 45 minutes.

After some high quality hors d'oeuvres the audience sat down to

listen to Dr Pierre Lafaye de Micheaux of the School of Mathematics

and Statistics, University of New South Wales, deliver a presentation

titled "A notion of depth for curve data", which uses ideas from

Princeton statistician John Tukey from the 1970s concerning, for

example, half-space depth for point clouds.

Pierre's research in this area is motivated by data from the

Older Australian Twins Study. There is strong evidence that the

quality of brain fibres impacts quality of life and, therefore,

high quality analyses of brain fibre data is important. The

data are curves in three dimensions so Pierre has had to

extend data depth ideas to this setting. The presenter made

excellent use of three dimensional graphics to visualise

the data and explain the depth concepts.

Another key idea was the concept of parametrised curves

and this involved some elegant geometry-type mathematics

including, of course, the Frechet metric.

After a theoretical exposition Pierre demonstrated his

breadth as a statistician by telling everyone about his

co-authored R package named curveDepth.

Apart from data from the Older Australian Twins Study the

methodology was applied to cyclone paths in the Gulf of Mexico.

The new methodology leads to better detection of outlier cyclones

and better confidence regarding regions at risk. For the brain

fibre applications, an upcoming challenge is to go from data

on 68 brains to 20,000 brains.

Matt Wand

University of Technology Sydney

The Statistical Society of Australia now has an official statement on climate change and the urgent need for action. You can read the statement here. The statement wording was made in collaboration with our Environmental Statistics Section and other experts, with thanks to David Warton. The statement was approved by the Executive committee in October 2019.

I am happy to discuss this any time.

Adrian Barnett, President

The Young Statisticians Conference 2019 (YSC19) – a forum designed for students and early-career statisticians – was held during 1-2 October in Manuka Oval, Canberra. Presentations were given by invited keynote speakers Teresa Dickinson (ABS), Calvin Hung (QuantumBlack), Alison Presmanes Hill (RStudio), and Margarita Moreno-Betancur (VicBiostat). Each elucidated the current trends in their respective fields, but laced their presentations with stories of experience, highlighting their own career-related journeys while giving little nuggets of wisdom to our enthralled audience. YSC19’s Careers Panel consisted of Louise Ryan (UTS), Teresa Neeman (ANU), Warren Muller (CSIRO), and Smitha Ramaswamy (Teachers Mutual Bank/UniBank). The invited speakers and panellists provided a much-needed insight for YSC19 delegates, with each demonstrating the multiple and diverse career paths available for our young statisticians.

In addition to the excellent presentations by our seasoned professionals, the talent amongst our delegates became evident very quickly, with demonstrated statistical applications ranging from agriculture to computer science. Amid the plethora of excellent works, Sarah Romanes from the University of Sydney won the Louise Ryan Best Presenters Award, the first ever named award in YSC history. The second and third Louise Ryan Best Presenters Award was awarded to Sayani Gupta from Monash University and Laura Cartwright from University of Wollongong. Details of their presentations are available in the above-mentioned website.

A video competition was established for students/early-career researchers who were unable to attend YSC19 but wished to be a part of the action. First prize went to Tan Jing Yi Joshua (Singapore Management University) for his video titled ‘Remember.For.Me’. Joshua and his colleagues conducted surveys to evaluate the degree of misunderstanding regarding dementia among the Singaporean population. The runner-up was Mustafa Shaheer Hamid (Monash University). Mustafa’s video, titled ‘Employing Semantic Segmentation Using Neural Networks’, used the segmentation method on the data from the Large Hadron Collider (LHC) to remove noise from a particle shower, which results in more accurate energy measurements of the particles.

With hard work comes a little bit of play. The YSC19 dinner, held at the Kingston Hotel, was a hit. YSC19’s social media guru, Kylie-Ann Mallitt (UNSW), hosted a Wiki-a-thon. The objective: to create Wikipedia pages for influential women in statistics using our smartphones. SSA President Adrian Barnett (QUT) conducted a mini competition to draw a perfect normal distribution. The winners were Rushani Wijesuriya and Nicholas Tierney.

It was an incredibly productive, engaging, and entertaining conference, to say the least. Special thanks to SSA, particularly Marie-Louise Rankin, whose tireless efforts ensured the success of this amazing conference. We would like to extend our heartfelt gratitude to the keynote speakers, panellists, Adrian Barnett, and our YSC19 delegates. We hope to see you all at our next YSC conference.

Janan Arslan

 

What is open banking?

Over the last few years, open banking has been predicted to be the next big thing, giving everyone vastly increased control over how their financial data is used. And it’s here now.

The demand for easier access and better control over our money isn’t new. Recent attempts to bring about this change include the Facebook Payment functionality through Messenger (2015) and the PayPal subsidiary Venmo, which simplifies the process of paying and receiving money from your friends and family.

Introduced last year, the Comprehensive Credit Reporting scheme gives everyone a better indication of their credit history and borrowing power. But the introduction of open banking in Australia offers an even larger shift in how we access and control our financial data.

.Know more.

Having already gained traction in the United Kingdom, European Union, the United States and Singapore, open banking is set to revolutionise the way we manage our finances in Australia.

According to the Customer Owned Banking Association (COBA), “Australian consumers are set to gain unprecedented access and control of their personal financial data as the Australian Government and banks continue the roll out of the open banking system.”

So, what is open banking?

In simple terms, open banking is the free flow of financial data from one organisation to another. The data is controlled entirely by the consumer, which means everyday Australians will be able to decide which companies, third party Apps and financial institutions have access to their information and how it’s used.

All authorised deposit-taking institutions (ADI’s) have been part of the rollout. Other companies such as App developers and Fintechs (financial technology) will need to be accredited, adhering to strict security and privacy standards before they can join the open banking revolution.

While the full range of possibilities are still being explored, some early examples of how open banking can benefit you include:

·        Comprehensive data sharing with a range of budgeting apps to help you keep on top of your spending

·        More complete data sharing with accountants and financial planners to improve the quality of advice and assist in the preparation of your tax return

·        Full portability of your transaction history when opening an account at a different bank, creating a lifelong financial imprint

·        A more complete understanding of your financial data, helping you to make informed financial decisions and find the products and services that are most suitable for you

·        Improved oversight to help banks and other financial institutions create products and services that are more suited to your specific needs

Improved privacy and security

Legislation has recently been passed through government requiring all financial institutions to make your credit and debit card, and deposit and transaction data available. Mortgage and personal loan data will follow in 2020, with data for all banking products being made available by 2021.

This staggered approach is designed to give all affected organisations time to implement policies and procedures so that consumer data is handled correctly and in line with the new, much stricter security guidelines. The recent Consumer Data Right gives consumers the option to choose how their data is shared, and with who. Only organisations that you authorise to access your data will be able to do so, and only for the purpose you specify.

Now that it’s here, do we actually want it?

While certain sectors have welcomed the idea of open banking, a 2016 survey commissioned by COBA found that only 42 per cent of everyday Australians are interested in the benefits of open banking. Nearly a quarter of those surveyed were opposed to the idea, although the younger demographics are much more receptive.

With data privacy dominating the news in recent times with Facebook and others being under intense media scrutiny over leaks and data breaches, there’s little wonder that so many people are sceptical about anything that proposes the free flow of personal information.

Because of the increased access to our own financial data and greater competition among established financial institutions, COBA has spoken out in favour of open banking. COBA has stated that the adoption of open banking by customer-owned banks is “a natural fit” because of the industry’s customer-first approach to developing and delivering financial products and services.

UniBank is a division of Teachers Mutual Bank Limited ABN 30 087 650 459 AFSL/Australian Credit Licence 238981.
Membership eligibility applies to join the Bank. Membership is open to citizens or permanent residents of Australia who are current or retired employees, students and graduates of Australian universities or family members of members of the Bank.

UniBank is a Silver Sponsor of YSC2019.

In the ACEMS Public Lecture Series on October 23 at 6pm, the NSW Chief Scientist and Engineer, Prof. Hugh Durrant-Whyte, will be presenting a talk describing a number of efforts to apply Bayesian methods to the problem of minerals discovery and characterisation. Click here for more details and to register.

The symposium on Data Science and Social Good will be held at QUT, Brisbane, on Friday, November 29. The purpose of this symposium is to promote the merger of data science and social good, share success stories, discuss challenges and potential solutions, extend networks, and explore directions for new research. Click here for more details and to register.

Bayesian Logistic Regression in Practice, using R or Autostats 
Clair Alston-Knox, Samantha Low-Choy, Daniela Vasco

Location: Griffith University, Gold Coast Campus, Southport

This is a two-part series of workshops:  

Part 1: Understanding and communicating the meaning of logistic regression, 
including elicitation of regression coefficients for Bayesian priors
Presenters: Samantha (lead presenter & developer), Daniela and Clair (co-presenters, co-developers) 
Date: Tuesday, 19 November, 2019; Gold Coast campus, Griffith University

Part 2: Inference for logistic regression – classical issues and Bayesian solutions
Presenters: Clair (lead presenter & developer), Samantha (co-developer) and Daniela (tutor)
Date: Thursday, 28 November, 2019; Gold Coast campus, Griffith University

Further details including abstracts will be posted at
https://www.griffith.edu.au/research/research-services/researcher-education-development/workshop-calendar

Cost:

$150/day for students outside Griffith; $200/day for members of SSA or ASBA; $250/day for others

South Australia Branch Meeting, August 2019

The speaker for the August meeting of the SA Branch was a Statistical Consultant Dr Kathy Haskard. Her forty-year career has included statistical consulting work in agriculture, ecology, environment, fisheries, and industry. In particular, Kathy worked for Australian Commonwealth and State governments, Universities, CSIRO, and Australia’s largest commercial specialist statistical consulting company, in locations spanning five Australian states and territories.

Kathy’s talk “Colourful Statistics – Monitoring Indigenous rock engravings on the Burrup Peninsula WA ” is a piece of her consulting work done at Data Analysis Australia jointly with Dr John Henstridge in 2017. Kathy was highlighted a number of statistical issues, practical difficulties and interesting learning from a statistical review of monitoring of ancient indigenous rock art annually over 13 years, including quantifying colour, spectrophotometers (L*a*b* scale) and spectrometers (reflectance spectra), BACI designs, varying design parameters, calibration, structured longitudinal multivariate data with irregularities, and mixed effects models.

In particular, she highlighted the following outlines in her talk

• Setting the scene - Burrup Peninsula Petroglyphs.  Archeologist Dr Mulvaney has published his research of just some of what is estimated to be over one million rock carvings, known as petroglyphs, on the Burrup Peninsula just 15 minutes' drive from the Pilbara mining town of Karratha
• Monitoring by CSRIO Mineral Resources annually from 2004. Seven petroglyphs have been selected with three additional industry sites, 3 spots at each site.
• Describing, measuring and quantifying colors by Spectrophotometers and ASD near-infrared spectrometer
• Data were hierarchical in nature and explored by dendrograms and Before–After Control-Impact (BACI) design and dimension reduction technique.

The detail of her talk and particularly quantifying colours by spectrophotometers (L*a*b* scale) and spectrometers (reflectance spectra) can be found by contacting kathyhaskard@gmail.com.

A dinner was held right after the meeting at Jasmin Indian Restaurant, 31 Hindmarsh Square, Adelaide.

By Shahid Ullah

South Australia Branch Meeting, September 2019

The speaker for the September meeting of the SA Branch was Peter Josef Kasprzak. The talk mainly discusses his master’s work at the Biometry Hub, Adelaide University which looks at the use of sampling methods in agriculture trials. The ultimate aim is to make robust sampling methods available to end users such as farmers who may have little computational experience. However, first there has to be a proof of concept for a semi-automated sampling protocol in a typical agricultural scenario. The motivating scenario was a seed trial where the emergence of faba bean seedlings was estimated. Automatic data collection using drones was explored and computer vision and machine learning techniques for image processing was compared in a sampling context.

A data collection protocol for sampling is not always conducted in agricultural experiments. An exploration of the literature found that often papers offered no justification for sample sizes used. This exacerbates the issue of non-reproducibility of experiments highlighted by Baker (2016) where more than 70% of researchers failed to reproduce another scientists experiments and 50% failed to reproduce their own.  It’s clear that an unbiased random sampling protocol is needed to ensure good study design.

Mcintyre (1952) introduced the ranking set sampling (RSS) method (without theoretical justification). RSS compares the units before the final selection which increases the structure on the measured data, without analysis of all the sampled units (Ozturk and Wolfe, 2000). In common agriculture distributions RSS performs better than other sampling methods for example simple random sampling SRS. Takahasi and Wakimoto (1968) showed the theory behind RSS showing the variance of RSS will always be less than SRS. RSS has the advantage that ranking can be done with auxiliary variable which are highly correlated and easily gathered variables. While the infinite paradigm of RSS is well mapped the finite paradigm is not.

In a field trial scenario the sampling paradigm is finite. In finite sampling there are 3 different levels of replacement. No items returned, all items returned and all items returned except for the selected unit. Simulation is an obvious tool to determine what sort and level of sampling should be undertaken. R Shiny was used to create a web based app which performs a simulation study based on different sampling protocol. Simply load in some data and simulations can be run in real time which robustly select the best sampling protocol. The app outputs a csv file with positions in the field to be sampled.

So how can this be implemented in the faba bean emergence trial? A drone was used to collect images of a field where faba bean seedlings were emerging at the plot level. Manual and automated flights were undertaken with image stitching of footage needed before processing for prediction of individual seedlings could be undertaken. Microsoft ICE and webODM, both free programs, used to stitch the smaller images taken by the drone together to form a composite image were compared.  webODM proved to be a superior program for handling changing conditions, but required an intermediate level of computing knowledge to use.  ICE was very simple to use, but was far more susceptible to changing conditions, and blurred joins in the final composite. With the final composite image, two approaches for processing the image and predicting the seedling numbers were compared-Computer Vision (CV) and Neural Networks (NN).  Python programs for CV and NN were used to obtain an estimate per pre-specified grid area in the field, which was then used in the sampling app as a ranking variable for the RSS protocol.  The study found that CV is easier to use but NN were superior as long as there is an adequate training set size which is diverse. In CV there are limited options to deal with false positives in high level stubble and potential over fitting can occur in NN with a greater level of work required to create the training data set. Issues such as seedling stress, nutrient and water stress effect the colours that you get in CV but NN can sort this out. An ideal solution is to use both in conjunction- start with CV to propose positive candidates and then use NN to disqualify false positives to leave only true hits to improve ranking.

The Shiny app sampling coordinates can be matched to the photographic coordinates. This means that images collected and processing by NN can be mapped to sampling coordinate position.  Pete presented results from the faba bean study which compared different finite sampling scenarios under SRS and RSS. The results showed that RSS was superior across all finite sampling replacement scenarios with relative efficiencies (variance SRS/variance RSS *100) between 143.1-216.1.

McIntyre, G. A. (1952). A Method for Unbiased Selective Sampling,Using Ranked Sets. Australian Journal of Agricultural Research,(3):385-390.

Ozturk, O. and Wolfe, D. A. (2000). Optimal allocation procedure in ranked set sampling for unimodal and multi-modal distributions. Environmental and Ecological Statistics, 7(4):343-356.

Takahasi, K. and Wakimoto, K. (1968). On unbiased estimates of the population mean based on the sample strati_ed by means of ordering. Annals of the Institute of Statistical Mathematics, 20(1):1-31.

The detail of his talk can be found by contacting peter.kasprzak@adelaide.edu.au. A dinner was held right after the meeting at Sukhumvit Soi 38, 54 Pulteney Street, Adelaide.

By Helena Oakey