Join our lab this summer! We are currently recruiting a senior undergrad student to help develop and evaluate a model of seafood fraud. We’re looking for students who are curious to expand their statistical and computational knowledge and experience.
Our lab is an interdisciplinary lab comprised of undergraduate and graduate students, as well as community partners, post-doctorates and other academics. The successful applicant will join a team of folks who are trying to better understand seafood fraud.
Details about the position are provided below.
Title: Bayesian Modelling of Seafood Fraud in the Canadian Supply Chain
Proposed Start Date: May 8, 2023 (negotiable)
Brief Outline of Proposed Research Project:
Seafood fraud is a mounting issue facing society today, where both economic loss and substantial harm to fish populations are immediately felt. One aspect of food fraud pertaining specifically to the Canadian market is seafood product mislabelling. Estimates suggest mislabelling rates of 50% and higher throughout the supply chain. Further, both the “when” and “where” of consumer product mislabelling remains largely unknown. Species mislabelling manifests at all times of year and involves many key stakeholders. Strong correlations also exist with species stock abundance, market price, and conservation status, complicating reliable fraud prediction.
The present work seeks to address the “when” aspect of seafood species mislabelling by developing a novel Bayesian hierarchical binary logistic time-series regression model in an effort to forecast the time-of-year of fraudulent behaviour for several key species of conservation concern commonly mislabeled at various points within the supply chain (e.g., within fisheries, supermarkets, and restaurants). Our model will greatly aid regulatory agencies in targeted and timely product testing in an effort to improve seafood authenticity and traceability within Canada. Part of the work will involve revisiting a controversial study by Stawitz et al. (2016) who employed Generalized Linear Models (GLMs) along with model selection using the Akaike Information Criterion (AIC). The study has garnered considerable criticism from researchers (DOIs: 10.1111/conl.12356, 10.1111/conl.12346, 10.1111/conl.12359), as well as justification from the original authors (DOI: 10.1111/conl.12384).
The current project, in collaboration with Food from Thought, along with researchers from the School of Computer Science and the Department of Integrative Biology, is ideally suited to a senior-level undergraduate student possessing interests in Bayesian statistics and computational algorithm development in the R and Stan probabilistic programming languages. Students should have successfully completed STAT*3240 (Applied Regression Analysis). Courses such as STAT*3510 (Environmental Risk Assessment), STAT*4000 (Statistical Computing), and/or STAT*4360 (Applied Time Series Analysis) would also be an asset. The end goal of the work is to produce an R package and/or a Shiny web app that will be uploaded to GitHub, submitted to the Comprehensive R Archive Network (CRAN), and hosted through an online web server for global uptake by the regulatory community and other interested parties.
The successful applicant will have strong background and knowledge of statistics, particularly regression modelling. Mastery of topics in areas such as time series analysis, environmental risk assessment, and statistical computing will be considered an asset. Additionally, the incumbent will possess strong programming abilities, preferably in the R programming language, as this project will involve the development of an R package and/or a Shiny web application for downstream use by regulatory scientists and other interested parties. The selected candidate will actively participate in weekly and/or biweekly lab meetings where they will present their research and ideas to the project investigators and lab group members, as well as contribute to drafting a manuscript to be submitted to a high-impact factor peer-reviewed academic journal.
Throughout the duration of the project the student will:
- become familiar with aspects of DNA barcoding, seafood market fraud, advanced regression techniques, Bayesian statistics, and Markov Chain Monte Carlo (MCMC) sampling, among other key topics
- become versed in retrieving and collating information from the primary literature and other sources to inform the desired model
- become proficient in coding and debugging within the R and Stan probabilistic programming languages through the RStudio integrated development environment (IDE)
- hone oral communication skills through presenting work and ideas at lab and other research meetings
- improve written communication skills through participation in the drafting of an academic peer-reviewed manuscript based on the completed work
Supervision, Reflection, and Feedback:
- The successful candidate will meet with the interdisciplinary research team at least weekly to review progress, identify challenges, and develop solutions.
- The successful candidate will gain experience working with and learning from knowledge domain experts from academia as well as from the community (i.e. government agencies, industry). Prior to meeting domain experts external to the academic community, the successful candidate will work with the research team to familiarize themselves with the partners, and to learn best practices and knowledge mobilization/transfer skills for interacting with non-academic experts and other community partners.
- The successful candidate will meet with the research team once per month to reflect on the project, the process, the experience, and how this information might impact how they work in the future.
For more information, feel free to contact me. If you are interested in applying for this position, please do so here. You may need to log into the Experience Guelph website first.