Past Meetings - QLD Branch
Wednesday 30 November 2011
Christmas Event --- Quiz Night
Time: 5:00pm– 7:00pm.
Location: Gibson Room, Level 10, Z Block, Gardens Point Campus, Queensland University of Technology. Click here for directions, and here for transport and parking.
Quiz Master: Adrian Barnett
Come and join us for a night of drink, food and fun. To celebrate our last meeting for the year the Queensland Branch of SSAI will be holding a statistical themed quiz night. You don’t need to form a team yourself (unless you wish to), just turn up and teams will be arranged on the night. A prize will be awarded to the winning team. Finger food and drinks will be provided and non-members are welcome.
Friday 5 August 2011
Young stats event
Time: 4:30pm– 7:00pm.
Location: The Gardens Point Guild Bar, Level 3, L Block, Gardens Point Campus, Queensland University of Technology.
All career young statisticians are invited to the Gardens Point Guild Bar for a young statisticians networking event. There will be some pizza and drinks provided free of charge.
You need not be a current member of the Statistical Society of Australia to attend.
If you’re a student (undergraduate or postgraduate) studying some statistics, work in areas related to statistics or have an interest in statistics and are within 5 years (or so) of commencing your career then please come along and meet your fellow career young statisticians.
Thursday 14 July 2011
Time: 4:30pm – 5:15pm.
Speaker: Ross Darnell, CSIRO.
Title: The data-reality gap: sampling issues in measuring biodiversity.
Followed by Annual General Meetings of The Statistical Society Of Australia Inc.
Time: 5:15pm – 6:15pm.
Click here for the agenda.
Location: St John's College, St Lucia Campus, University of Queensland. Click here for the location, and here for transport information.
Wednesday 15 June 2011
Speaker: Prof Joe Whittaker, Mathematics and Statistics, University of Lancaster.
Title: Divergence weighted independence graphs for the multivariate analysis of survey data
Time: 5:00pm – 6:00pm, followed by dinner at The Caddy Shack Cafe.
Location: Room 015, Level 0, Public Health Building, Herston Rd, Herston Campus, University of Queensland Maps: Map one or Map two.The Royal Children's Hospital bus stop is close by.
The analysis of survey data, collected on a set of response variables defined over a finite population, benefits from a bird's eye view of their inter-relationships and in particular, of their strengths. This overall analysis should highlight those variables that strongly modify the conditional distribution of another variable, and by contrast, should indicate those which have little affect. The weighted graph based on divergence measures of independence strength calculated from the population fulfils this purpose. The bootstrap gives pointers for establishing how large conditioning sets may become while retaining stable estimates.
Biography: Professor Joe Whittaker is a long term member of the Department of Mathematics and Statistics, Lancaster University. He is a fellow of the Royal Statistical Society, a past member of its Council, of its Research Section, and past chairperson of its Multivariate Study Group. He has been a member to the Board of Directors of the European Region of the International Association for Statistical Computing. He has held visiting Professorships in the Department of Statistics at Colorado State University and at the University of Chicago. His main research interests lie in graphical modelling. Previous courses have been sponsored by the ESRC, and has been given in England, Belgium, Holland, France, Italy, Spain, Switzerland, Brazil and the USA. He is the author of Graphical Models in Applied Multivariate Statistics, published by Wiley.
Monday 30 May 2011
Speaker: Associate Professor Dr Noor Azina Ismail, Department of Applied Statistics, University of Malaya.
Title: Identifying and Predicting Students’ Attitudes toward Mathematics Using Latent Class Analysis
Location: QUT, Gardens Point Campus, Room 413, Level 4, Z Block.
Time: 5:00pm – 6:00pm.
Abstract: Latent class analysis (LCA) is a popular tool in identifying latent groups of related cases using a set of categorical or continuous observed variables. Since the data structure in this study includes students nested in school, a traditional LCA that assumes observations are independent of one another is not suitable to analyse this type of data. As such, a multilevel latent class analysis (MLCA) is needed. Latent classes of students’ attitude toward mathematics among 4466 Malaysian students in the 8th grade are considered. The best model comprised of three level 1 latent attitude and building on this three-class level 1 solution, a model that take into account for the nested structure of the data is specified. Several covariates at the student level, such as gender, parental influence and mathematics achievement are also considered in this study.
Biography: Noor Ismail is a research fellow under the endeavour award program and a visiting academic at the Mathematical Science Discipline, QUT. Her recent interests are focused on the latent variable modelling, multilevel analysis and structural equation modelling.
Monday 18 April 2011
Interrupted Time Series: What, Why and How – An Example from Suicide Research
Speaker: Karen Smith, Leicester Clinical Trials Unit, University of Leicester.
Time: 5:00pm – 6:00pm, Refreshments from 4:45pm.
Location: Room 413, Level 4, Z Block, Gardens Point Campus, Queensland University of Technology. Links to: Map; Public Transport and Parking
Karen's slides are available here.
It's well known that a well-designed randomised controlled trial (RCT) gives the best level of evidence of a healthcare intervention. But there are instances where it is impractical or impossible to conduct an RCT. Interrupted time series gives one of the strongest quasi-experimental approaches, useful when there is change at a particular point in time, and offers an improvement over the before vs. after analyses that are often conducted. However this approach is not widely known or used. In this talk an overview will be given of what interrupted time series is, why and when it should be used and how the analysis might be done. An example will be discussed to illustrate the approach based on work conducted with the Centre for Suicide Research at the University of Oxford.
Biography: Karen is the Principal Statistician at the University of Leicester Clinical Trials Unit, where she leads the statistical input and has responsibility for the design and conduct of clinical trials across all disease areas. Previously she headed the NHS Statistical Support Team at the Centre for Statistics in Medicine in Oxford, with a remit for both clinical trials and observational studies. Karen has worked at both Lancaster and Liverpool Universities as a medical statistician as well as spending some years in the pharmaceutical industry where she gained a sound understanding of the drug development process, as well as the difficulties this entails. She has also provided statistical support to research students at Coventry University across disparate disciplines, where her previous life as a teacher proved invaluable and a long-standing interest in modes of support for non-specialists was consolidated. Her current research interests are largely focused around clinical trials, in particular clinical trial design including appropriate methods of treatment allocation and the difficulties in cluster RCTs, repeatability and reliability studies and missing data problems.
Tuesday 8 March 2011
Queensland Branch Annual General Meeting
Time: 4:45pm – 5:15pm, Refreshments from 5:00pm.
Location: Room 320, Building 68 (Chemistry Building), St Lucia Campus, UQ.
Time: 5:15pm – 6:15pm, Refreshments from 5:00pm.
Location: Room 320, Building 68 (Chemistry Building), St Lucia Campus, UQ. Map; Transport and parking
While single processor computing power has grown rapidly over time, data collection has kept pace to the point where a single processor may no longer suffice with the massive amounts of data available. Parallel computing harnesses the power of multiple processors, either within a desktop or on a computing cluster, in order to increase the speed of analysis. Graphics Processing Units (GPUs) take parallel computing to the next level. Originally developed to produce high-level 3D graphics for video games, GPUs have recently been converted to use in scientific computing. A single GPU contains hundreds of computing cores which can result in orders of magnitude increases in speed for certain applications. I will give a brief overview of GPUs and how they work, and then present examples of their applications, including a case study in statistical genetics of my own attempt to work more quickly.
Followed by dinner at a nearby restaurant.
Monday 13 December 2010
Professor Denise Lievesley
Time: 4:00pm – 5:00pm (drinks and nibbles afterwards).
Location: Change of venue: Room S402, Level 4, Building 24, St Lucia Campus, UQ.
Professor Lievesley is one of the country’s leading social statisticians, who has campaigned for evidence to be used as the basis for the development of sound public policies within the UK and more widely. Having enjoyed a distinguished career, which has included the posts of founding Chief Executive of the English Information Centre for Health and Social Care; Director of Statistics at UNESCO –where she established its new Institute for Statistics – , and Director of the UK Data Archive (and simultaneously Professor of Research Methods in the Mathematics Department, University of Essex), most recently Professor Denise Lievesley was a special advisor at the African Centre for Statistics of the UN and was based in Addis Ababa.
Professor Lievesley’s various roles have led her to work with ministers, ambassadors, senior civil servants and officials of international agencies, for which she has established a reputation for upholding the principles of professional integrity, policy relevance and methodological transparency. Throughout her working life, Professor Lievesley has been committed to protecting the integrity of official statistics and to ensure that they remain free from political influence.
please RSVP to Cassie Mellor (ISSR) at c.mellor@uq.edu.au for this Public Lecture if you can attend.
Friday 26 November 2010
Queensland Branch Christmas Bash
Time: 5:00pm – 6:00pm, Refreshments from 4:30pm.
Location: Owen J Wordsworth, Level 14, S-Block, Gardens Point Campus, QUT.
Directions: http://www.qut.edu.au/about/location/
Transport and parking: http://www.fmd.qut.edu.au/campus_services/parking/
Come and join us for Christmas nibbles and drinks from 4:30pm before SSAI Queensland Branch presents a selection of entries from the “Funny Stats Competition”. Last minute entries are welcome at the event. Finger food and drinks will also be provided following the presentation. Non-members are welcome.
To see the top three entries click here..
Tuesday 14 September, 2010
Young Statisticians Event: Career Information
Time: 4:30pm – 5:30pm
Location: Room 520, Level 5, O Block (Mathematics Building), Gardens Point Campus, Queensland University of Technology Directions: click here. Transport and parking: click here.
Speakers:
- David Priest, Queensland Investment Corporation
- Rachael Wills, Queensland Health
- Kristen Gibbons, Mater Hospital
This event is likely to be of interest to undergraduate students studying some statistics and anyone interested in entry level statistics positions, although all are welcome to attend.
Some young mathematics and statistics graduates will speak on their current statistical roles since graduating. There will be opportunity to ask questions to the speakers and socialise with other young statisticians. Refreshments will also be provided free of charge.
You need not be a current member of the Statistical Society of Australia to attend.
Tuesday August 10, 2010
Clinical Monitoring: An Evidence-Based Approach
Time: 5:00pm - 6:00pm (Refreshments 4:45pm)
Non-members are welcome. No RSVP required.
Location: The Hall (Z2-226), Musk Avenue, Kelvin Grove Campus, Queensland University of Technology; Google maps; Lots of parking bays on Musk Avenue or Carraway Street. There's also a car park at the top of Musk Avenue near Kelvin Grove Road.
Speaker: Prof Paul Glasziou, Centre for Research in Evidence-Based Health Care, Bond University
Monitoring and management of long term illness is an important and costly element of health care. It accounts for 80% of GP consultations, and a substantial proportion of laboratory testing. However, despite the substantial work and costs that it entails, monitoring has been neglected as an area for research. For example, despite poor evidence of effectiveness of self-monitoring in type 2 diabetes, the cost of monitoring strips alone in 2002 in the UK was £118 Million per year: larger than the expenditure on oral medications for diabetes. With clear evidence now that daily glucose monitoring is not warranted in type 2 diabetes, the rates are now declining. Similarly the clinical monitoring for several other conditions - including lipid management, blood pressure management, and asthma - do not align with the current evidence. Despite clinical efforts in monitoring, in many patients chronic disease is poorly controlled. For example, in a UK study before the new GP contract, only 14% of 21,024 newly diagnosed patients with hypertension had met target blood pressure after 12 months, and among treated patients about 40% of INRs are outside target ranges, compared with the ideal of 5%. Hence development and usage of optimal methods for clinical monitoring have the potential to improve care and reduce costs. We suggest [1] that optimal monitoring should be considered in five phases: (1) pre-treatment, (2) initial titration, (3) maintenance, (4) re-establish control, and (5) cessation. Current randomised trials of monitoring showed a mixed pattern of effects, but few appear to have optimized monitoring protocols prior to the trial, or consider the need for optimisation of each of these stages. This should include consideration of measurement variability, patient self-monitoring, and the patient-clinician-laboratory interfaces [2].
References:
[1] Glasziou P, Irwig L, Mant D. Monitoring in chronic disease: a rational approach. BMJ. 2005 Mar 19;330(7492):644-8.
[2] Glasziou PP, Irwig L, Aronson J. Evidence-Based Medical Monitoring: Principles and Practice. Blackwells Books, 2008.
Biography:
Professor Paul Glasziou MB BS, PhD, FAFPHM, FRACGP is currently the Director of the Centre for Research in Evidence-Based Health Care at Bond University, and Professor of Evidence-Based Medicine in the Department of Primary Care at the University of Oxford, and also continues work as a part-time General Practitioner. He holds an NHMRC Australia Fellowship to undertake research on the causes and cures for the evidence practice gap.
Tuesday July 27, 2010
Model Based Geostatistics
Time: 5:00pm - 6:00pm (Refreshments 4:45pm)
Location: Owen J Wordsworth Room, Level 12, S Block; Gardens Point Campus, Queensland University of Technology; Directions; Transport and parking
Speaker: Dr. Archie Clements, School of Population Health, University of Queensland
Geostatistics is an interpolation method that uses information from a limited set of locations to predict values at all, or a subset, of locations in a study area. The fundamental basis of geostatistics is spatial autocorrelation, where observations from nearby locations tend to be similar. While traditional geostatistics has advantages over other interpolation methods, it is not ideally suited to non-Gaussian response data and has limited utility in estimating prediction uncertainty. To address these issues, Diggle and colleagues developed model-based geostatistics (MBG). This embeds geostatistics in a generalised linear modelling framework. Via the GLM architecture, a flexible range of different types of response variable can be accommodated. Additionally, MBG employs a Bayesian method of statistical inference, thereby providing an intuitive interpretation of parameter uncertainty whilst explicitly modelling spatial autocorrelation and, most importantly, allowing a formal expression of uncertainty in the prediction estimates.
The application of Markov chain Monte Carlo simulation (MCMC) for model fitting makes the considerable computational demands of integrating functions over multiple dimensions achievable. MBG has increasingly been used in epidemiology to predict the distribution of diseases and associated uncertainty. I illustrate the applicability of MBG in providing an evidence base for control programme planning and resource allocation using the example of tropical parasitic diseases.
Friday June 25, 2010
Systematic Reviews and Meta-Analyses: Half-Day Meeting
Time: 9:00am-11:30am, followed by lunch.
Location: Seminar room, IHBI Building Kelvin Grove Campus, Queensland University of Technology Please see this webpage for directions.
Speakers:
- Prof. Jenny Doust (Bond) “Doing a systematic review with the Cochrane Collaboration”
- A/Prof. Suhail Doi (UQ) “Combining heterogeneous studies using the random-effects model is a mistake and leads to inconclusive meta-anlayses”
- Dr. Adrian Barnett (QUT) “Incorporating study quality into meta-analyses”
- Dr. Henry Zheng (QUT) “The effect size and dose-response of walking interventions to prevent CHD: Meta-analysis”
This meeting is free and all researchers conducting a systematic review or meta-analysis (not just statisticians) are welcome to attend. Please register with Adrian for catering purposes (a.barnett@qut.edu.au). After the four speakers we hope to have a lively question and answer session. If you have any questions about systematic reviews or meta-analyses that you’d like to be answered by our four experts then either e-mail Adrian in advance or bring your question on the day.
Friday May 21, 2010
Young Statisticians Networking Event
The Gardens Point Guild Bar, Level 3, Y Block, Gardens Point Campus, Queensland University of Technology.
All career young statisticians are invited to the Gardens Point Guild Bar for a young statisticians networking event. There will be some pizza and drinks provided free of charge. You need not be a current member of the Statistical Society of Australia to attend.If you’re a student (undergraduate or postgraduate) studying some statistics, work in areas related to statistics or have an interest in statistics and are within 5 years (or so) of commencing your career then please come along and meet your fellow career young statisticians.
ORDINARY BRANCH MEETING: Tuesday 24 November 2009 at QUT, Gardens Pt
Christmas Event 
Frequentist vs. Bayesian Debate
Speaker: Frequentists vs. Bayesians
Date: Tuesday 24th November, 2009
Time: 5:00pm – 7:00pm
Location: Owen J Wordsworth Room, Level 12, S Block
Gardens Point Campus, Queensland University of Technology
http://www.qut.edu.au/about/location/
Details:
Come and join us for Christmas nibbles and drinks from 5:00pm before getting your ringside seats
to the statistical event of the year. First round commences at 5:30pm. Post-match celebrations
follow, including finger-food and drinks.
Adjudicator
Referee Robert Reeves
Frequentist Corner (THE WINNERS!)
Roundhouse Rodney Wolff
Adrian Boom Boom Barnett
vs.
Bayesian Corner
Tony the Tornado Pettitt
Kerrie Lean Machine Mengersen
ORDINARY BRANCH MEETING: Tuesday 13 October 2009 at UQ, St Lucia
SPEAKER: Dr Paul Jackway, Principal Research Scientist in the Quantitative Image Analysis group, Mathematical and Information Sciences, CSIRO, QLD
TITLE: Ten Things to Do With a Dirty Picture
TIME: 4:00- 5:30, Tuesday 13th October 2009
VENUE: Room 67-442 (67 = Priestly Building; Room 42 on level 4), University of Queensland, ST LUCIA
Map: see www.uq.edu.au/maps/index.html Transport and parking: www.uq.edu.au/maps/directions.html
DINNER: Members and guests are invited to join the speaker at a nearby restaurant.
Abstract
This talk will attempt to convey some of the breadth of contemporary image analysis without getting too bogged down in the detail; broad rather than deep it will target the statistician who wants to forget about numbers and algorithms and theorems for an hour and look at the big picture(s).
From the vantage point of nearly 2 decades in the field of computer image analysis, I will talk about some of the things I have found to be “cool” possibly including: Watersheds, Granulometries, Snakes, Mathematical Morphology, Scale-Space, Top-hats, Texture Measures, Zero-crossings, Histograms, Edges… The starting point will be the difference between a picture and an array of numbers.
Those wondering about the title should first Google “Julian Besag” “Dirty Pictures”.
Biography
Paul studied electronics at RMIT in the early 1980's followed by applied statistics, also at RMIT, in the later part of that decade. During this time he was employed by Telstra first as a product engineer and later in business planning and support performing business data integration analysis. In 1991 he moved to Brisbane to undertake graduate studies at QUT in mathematics and computer image analysis and was awarded a PhD in 1995.
Following his PhD submission, Paul obtained a research fellowship in the Cooperative Research Centre for Sensor Signal and Information Processing (CSSIP) based at the University of Queensland. For the next eight years at CSSIP, Paul led a research group working on automated cytology focussing on the automation of Pap smear cytology for cervical cancer screening. In 2002 Paul accepted a Principal Research Scientist position in the Quantitative Image Analysis group within the Mathematical and Information Sciences Division of the Commonwealth Scientific and Industrial Research Organisation (CSIRO). Since then he has contributed to a range of projects and has undertaken general image analysis research while leading the group’s strategic push into automated microscopy and photography.
ORDINARY BRANCH MEETING: Tuesday 1 Spetember 2009 at QUT, Gardens Point
SPEAKER: Greg Waite, Principal Data Analyst, Organisational Performance and Strategy, Department of Communities, Qld Government.
TITLE: Use of a data set for in depth analysis of an administrative dataset for informing evidence based policy.….
TIME: 4:00- 5:30, Tuesday 1st September 2009
VENUE: Room 520, level 5, O Block (Mathematics Building), Gardens Point Campus, QUT
Maps, Transport and parking: http://www.qut.edu.au/about/location/
DINNER: Members and guests are invited to join the speaker at a nearby restaurant.
Format of Meeting *
This is a new interactive style format- where Greg will present the variables and his needs from the data. We hope post grads and honours students who are interested in real life large datasets may come along, to see an interactive discussion of how the data will be summarised, and what questions may be needed from policy makers. ( If you are interested we may be able to give the variables names and intentions of analysis a few days ahead- so you can develop your own questions of interest)
The participants can assist in suggesting novel ways of analysing the data. This is a hands on practical application of – scoping of methods to a real social problem. This data set has a depth of information awaiting the right analysis and summary to inform policy.
* New Style - participatory
For more details of our next meeting, please see http://ssaqld.wordpress.com/
ORDINARY BRANCH MEETING: Tuesday 4 August 2009 at UQ, St Lucia
Queensland Stats Brain Challenge
Student Symposium and Quiz Night
SPEAKERS: TBA
VENUE: Room 67-442 (67 = Priestly Building; room 42 on level 4), University of Queensland, ST LUCIA
Map: see www.uq.edu.au/maps/index.html Transport and parking: www.uq.edu.au/maps/directions.html
The QLD branch of the Statistical Society of Australia invites all members and guests to our student symposium and quiz night. The night will showcase the work of some of our up and coming statisticians, and provide an opportunity for members to enjoy each other’s company and test their knowledge of statistics, with the inaugural Queensland Stats Brain Challenge!
Put your knowledge of statistics trivia to the test!
Program
5:00: Student Talks - 3 minutes and 1 slides! Prize for 2 Best Presentations - $25 gift voucher. Register now - email Rob Reeves r.reeves@qut.edu.au.
5:30: Pizzas, wine and cheese.
6:00 Quiz Register your team of three now, email r.reeves@qut.edu.au. Teams must provide a
set of 10 questions on statistics related trivia.
Prizes: Win bottles of wine and beer, as selected by our prize committee!
Come along and enjoy a night of fun, and enjoy seeing what the next generation of statistics professionals are up to!
Register now: RSVP by Friday 31st July to Rob Reeves r.reeves@qut.edu.au or 3138 2827 (BH).
ORDINARY BRANCH MEETING: 14 July 2009 at ABS, City
SPEAKER: Professor Chris Lloyd, Associate Dean of Research, Melbourne Business School, The University of Melbourne
TITLE: Statistical Blogging: The Fishing in the Bay Story
TIME: 3:45pm for 4pm talk, Tuesday, 14 July, 2009
VENUE: Training Room 1, Level 3, Australian Bureau of Statistics, 639 Wickham St, Fortitude Valley, Brisbane.
*ABS Office: take lift to Level 3 above Freedom Furniture (Homemaker Centre)
Google map here
DINNER: All are invited to join the group at a nearby restaurant.
ABSTRACT:
Prior to the web, professional peers were linked though societies and the conferences and newsletters they sponsor. Interactions were limited to members and specific times. The first professional forums that used the web were professional lists (such as anzstat) and discussion boards (such as radstats) where participation was less restricted and communication more immediate. During the past 5 years a new platform, the weblog, has emerged. The number of weblogs has doubled every 6 months, yet only a small minority are oriented towards particular professional communities, let alone statisticians.
In this talk I will talk about blogs in general and why I decided to establish one for the Australasian statistical community. Have you ever wondered how to set one up, how much time it takes to run and how I find material to post? Perhaps not, but I will tell you anyway. I will illustrate the advantages of the platform as I see it. Some of my favourite posts will be described in detail. I will finish by describing how FIB could be a much more effective space for our profession, with only a small but consistent input from my fellow statisticians.
BIOGRAPHY:
Chris Lloyd's recent research focuses on accurate inference for count data, including so-called exact methods, higher order asymptotics and bootstrap. His recent teaching is centred on business statistics for MBA students, and emphases hands-on computation, how to assess rather than ignore uncertainty, and how to combine several simple techniques to unpack complex business problems. The talk however is about the third strand of the so-called triple helix - knowledge transfer. For the past three years, Chris has run a weblog covering all areas of statistics and probability.
For more details of our next meeting, please see http://ssaqld.wordpress.com/
A Three-Day Course on Practical Bayes for Beginners, Brisbane
Presented by Kerrie Mengersen, Centre for Data Analysis, Modelling & Computation, QUT
26, 27 and 28 August 2009.
For more information, please click here.