Friday, November 19 – Saturday, November 20, 2021
CANSSI will be hosting a virtual event celebrating the work of our researchers, postdocs, and students. This two-day event, running Friday, November 19 and Saturday, November 20, will be a great opportunity for you to meet the national statistical and data science community, hear about career opportunities, and learn about different ways CANSSI can support your work. This event includes:
- A community town hall
- Presentations by CANSSI Collaborative Research Teams (CRT)
- A panel session
- Lightning talks
- Poster session for postdocs, grads and undergrad students
- A social/networking event
Participating in the poster session allows other academics and researchers to see your work. You’ll get a chance to network with faculty across the country during the poster session. Limited spaces available.
There will be lightning talks, which are short, 10 minute presentations. These talks offer a great opportunity to share your work with other academics and researchers, and to hone your presentation skills.
EDI Town Hall
The Statistical Society of Canada and CANSSI are organizing a townhall on their Equity, Diversity and Inclusion activities, policies, and programs. After short overviews of the programs, we will organize a community discussion. We are seeking evaluations, experiences, and new ideas from the community. The goal is to help ensure that our EDI activities and programs are community driven.
Day 1: Friday, November 19
(EST / PST)
|12:00 / 9:00||Welcome|
|12:15 / 9:15||CRT Presentations by Mahmoud Torabi and Changbao Wu|
|13:45 / 10:45||Break|
|14:00 / 11:00||CANSSI Program Overview – Meet the Directors|
|14:45 / 11:45||Lightning Talks|
|16:00 / 13:00||Break|
|16:30 / 13:30||EDI Townhall|
Day 2: Saturday, November 20
(EST / PST)
|12:00 / 9:00||Workshop: Innovations in Data Analytics and|
Data Science: What’s New at Statistics Canada?
|15:00 / 12:00||Break|
|15:30 / 12:30||Lightning Talks|
|16:30 / 13:30||CRT Presentations by Tianyu Guan and Ken Peng|
|17:15 / 14:15||Poster Session, Networking & Social Hour|
CRT Presentation Summaries
Sports Analytics Using Event Data and Tracking Data
Presented by Tianyu Guan
This presentation will first talk about a project in the National Rugby League (NRL) which aims to develop new methods for providing instantaneous in-game win probabilities. Besides the score differential, betting odds and real-time features extracted from the match event data are also used as inputs to inform the in-game win probabilities. Rugby matches evolve continuously in time and the circumstances change over the duration of the match. Therefore, the match data are considered as functional data and we propose methods to estimate the in-game win probabilities from the perspective of functional data analysis. I will then introduce our recent research on the strategy of playing defensively in soccer. The analysis is predicated on the assumption that the area of the convex hull formed by the players on a team provides a proxy for defensive style. It is assumed that smaller areas coincide with a greater defensive focus. With the availability of tracking data, the area of the convex hull can be obtained. Finally, I will outline the goals of our CRT.
About Tianyu Guan
Tianyu joined the Department of Mathematics and Statistics at Brock University as an Assistant Professor in 2020. She completed her BSc in Statistics from Jilin University in 2011. She received her MSc in Actuarial Science from Simon Fraser University in 2014 and then obtained her PhD in Statistics in 2020. Her research interests are sports analytics, functional data analysis and data science.
Applying Bayesian Methods to the Impulse-Response Modelling of Elite Middle-Distance Runner Performance
Presented by Ken Peng
The impulse-response (IR) model describes the relationship between athlete training history and performance. The model features five parameters, with two derived parameters providing context for interpretation in terms of exercise training. Despite some past successes, IR models are often poorly fitted. Here we describe a novel Bayesian approach to fit the IR model. We discuss the elicitation of informative priors, and justify the assumption that performance is multivariate normal distributed. MCMC via Gibbs sampling was used to sample the posterior. The method was applied to an international-class middle-distance runner, for which training was quantified as TRIMPi and performance as IAAF points achieved in a sanctioned race. The method produced well-constrained estimates of the five parameters, but the posterior intervals of the derived parameters were too wide to make reliable training optimization decisions. We conclude our approach could further improve the fit of the IR model.
About Ken Peng
Ken is a 2nd-year master’s student majoring in Statistics. Ken is a big fan of sports analytics and he finished third in the Young Investigator Award competition at the Sport Innovation Summit 2020. His master’s thesis includes the Bayesian implementation of the critical velocity model which is well-known in sports science, and he is going to be defending in the Fall.
A Review on Some Recent Advances in Infectious Disease Modeling
Presented by Mahmoud Torabi
Identifying disease clusters and spatial patterns of disease (from human/animal/plant) is important to inform policy makers, programs and interventions at both local and global scales. The recent emergence of the coronavirus as a global pandemic is one example of a critical public health threat that challenged management systems. The rapid spread of coronavirus across much of the globe is not well understood yet. Space-time patterns of spread span multiple scales due to complex disease ecological processes and biases from surveillance data generated from multi-jurisdictions with varying sampling protocols are real challenges. Our CRT members have spearheaded innovation in disease modeling by addressing several practical problems related to infectious diseases in environment and health by advancing statistical modeling techniques. In this talk, I will review some projects done by our CRT members in the context of infectious disease modeling.
About Mahmoud Torabi
Mahmoud is a Professor of Biostatistics in the Department of Community Health Sciences at the University of Manitoba. He is also an Adjunct Professor in the Department of Statistics and a scientist in the Children’s Hospital Research Institute of Manitoba (CHRIM) at the University of Manitoba. His main research areas are spatial statistics and small area estimation. He has received some provincial (Research Manitoba, CHRIM), and national funding (NSERC DG, NSERC Alliance, NSERC EIDM, CANSSI-CRT, CIHR) for his research as principal investigator (PI) and co-PI. He has published more than 65 papers in peer-reviewed statistics and health research journals, and served as a referee for over 100 papers. He has served the Statistical Society of Canada in various capacities including President of the Survey Methods Section.
Empirical Likelihood Inference for Non-probability Survey Samples Presented by Changbao Wu
In this presentation, we first provide an overview of two major developments on complex survey data analysis: the empirical likelihood methods and statistical inference with non-probability survey samples. We then propose new inferential procedures on analyzing non-probability survey samples through the pseudo empirical likelihood approach. The proposed methods lead to asymptotically equivalent point estimators that have been discussed in the recent literature but possess more desirable features on confidence intervals such as range-respecting and data-driven orientation. We conclude with some further discussions on issues and problems related to the topic.
About Changbao Wu
Changbao is Professor of Statistics in the Department of Statistics and Actuarial Science at University of Waterloo. His main research interests include design and analysis of complex surveys, resampling techniques, missing data analysis and causal inference, and data integration. He is Fellow of ASA, Fellow of IMS, Elected Member of ISI, and was the winner of the CRM-SSC Prize in Statistics in 2012. He has served on several editorial boards including CJS, JASA and Biometrika. He is the lead author of the book Sampling Theory and Practice (with Mary Thompson) published by Springer in 2020.
Innovations in Data Analytics and Data Science: What’s New at Statistics Canada?
This panel session will feature developments in data analytics and data science at Statistics Canada, including simulation methods and datasets, new research projects, available data resources, and innovative processes and practices to ensure analytic approaches are transparent and equitable.
EDI Town Hall Speakers
Bouchra Nasri is Assistant Professor at the School of Public Health of Univesité de Montréal. Her research interests are dependence modelling, time series, and more recently spatial modelling. She also authored and co-authored several R packages.
The main applications targeted by her research projects are related to climate change, public health, and infectious diseases modelling. Dr. Nasri is an associate director of the new infectious diseases network OMNI-RÉUNIS and co-lead of the data theme for the network. Her work is funded by multiple provincial and federal agencies. Her work is statistics is found mainly by NSERC and FRQNT. Her applied work in public health is found by multiple source such as CIHR, CReSP, ESPUM, PHAC and more recently MfPH and OMNI-RÉUNIS. Finally, Dr. Nasri is Chercheure boursière Junior 1 in Artificial intelligence in health and digital health
Don Estep, Scientific Director at CANSSI and Simon Fraser University Science Professor, is a mathematician and statistician who collaborates with scientists and engineers to develop algorithms and models for a range of problems in ecology, physics and engineering. He has applied his research to detect black holes, forecast hurricane storm surges, measure electromagnetic scattering and predict fusion reaction. By examining the behavior of complex multi-physics systems and subsequently creating computational and statistical models Estep can predict behavior, quantify uncertainty and ultimately mitigate hazards.