Joel Barnes » Data Scientist | Model Developer | Analytics

Joel Barnes

Data Scientist | Model Developer | Analytics

github.com/barnzilla linkedin.com/in/joel-barnes-5a80a449 Ottawa, Canada

Experience

Data Scientist

Statistics Canada
2020 - present Ottawa, Canada

• Develop and maintain an agent-based model (POHEM) by creating new sub-models that contain cross-sectional and longitudinal predictive algorithms, which forecast various health risk factors and outcomes (e.g., obesity, diabetes, dementia, cardiovascular disease) in Canadians from 2001 to 2050; used by the Government of Canada to inform policy decisions.
• Serve as lead for team members; help them troubleshoot various issues (e.g., predictive modeling guidance, simulation model compilation errors, Docker build issues).
• Develop and continuously test an R Shiny web-based lightweight UI for OpenM++ agent-based modeling that uses React and a key-value stores caching system (cachem); enables model developers to easily create/run model scenarios and download results.
• Developed and maintain an object-oriented R package (chms) that uses distributed processing to apply algorithms to national health data, reducing processing time by 10x.
• Built and manage an R package, ompp4aaw, that enables model developers across divisions within Statistics Canada to easily install, compile and deploy OpenM++ simulation models across machines/nodes with limited/no knowledge of Linux distributions and Kubernetes.
• Designed and developed an R Shiny web-based app, Data Dictionary Builder, that groups 10+ years of national survey metadata from flat files in a filterable/searchable format, making the lives of Statistics Canada analysts and external researchers simpler and easier by allowing them to identify variables and plan analyses more efficiently.
• Modeled Covid-19 cases in nine Canadian provinces on a weekly basis during the pandemic using compartment models via a co-authored R package, EpiSim, that forecasted cases one year into the future; deployed all models within a Microsoft Azure production pipeline using Apache Spark; tuned model performance (e.g., runtime and accuracy) on a weekly basis using updated case data, parallel processing and minimizing functions; models were used to help the Government of Canada forecast personal protective equipment demand and make evidence-informed procurement decisions.

Data Scientist

CHEO Research Institute
2008 - 2019 Ottawa, Canada

• Performed various modeling techniques on large datasets related to movement behaviours (physical activity, sedentary behavior, sleep) and health outcomes in children and youth. Analyses have been published in 70+ peer-reviewed publications contributing to an h-index of 40, which is equivalent to the research output of a full professor.
• Served as a full-stack web developer that designed, built and maintained five websites on LAMP stacks.
• Served as lead author and manager of contributing authors on seven ParticipACTION Report Cards on Physical Activity for Children and Youth (2011-2016, 2018), a knowledge translation product that has received one billion+ media impressions and that has an estimated collective ad value of $10 million+.

Freelance Web Developer

Barnzilla Web Development
2006 - present Ottawa, Canada

• Design, build and maintain custom websites for clients on virtual private servers (LAMP stacks); includes caching setup and website speed testing to ensure websites load quickly (e.g., scapps.org speed test)
• Develop and deploy custom code, incuding relatonal database modeling, for academic society clients to manage key academic conference functions including paper submissions, fee processing and registration;
• Develop custom web-based dashboards that reshape, filter and visualize data based on user behavior (e.g., Canada Grains Council Trade Data Online, Covid-19 cases in Canada).
• Have designed and developed 50+ websites for clients since 2006.

Skills

Modeling, Analysis and Visualization
R, Python, OpenM++, C++, SQLite, Shiny, Quarto

Builds, Deployment and Pipelines
Azure Data Factory, Azure Data Lake, Azure Databricks, Apache Spark, DigitalOcean droplets, Kubeflow, Kubernetes, Jupyter Notebook, Docker, Grafana

Stacks
Debian, Ubuntu, Apache, Nginx, MySQL, MariaDB, PHP

Frontend
HTML, CSS/SCSS, jQuery, JavaScript, DataTables, WordPress, React, Vue

Education

MSc in Kinesiology

University of Saskatchewan
Saskatoon, Canada