Research Data Analyst
The ASIST Project, led by neurosurgeon Dr. Christopher Witiw and Scientist Dr. Christopher Smith, is dedicated to advancing innovative AI-driven tools that enhance clinical workflows and improve patient outcomes.
Our team is currently looking for a Research Data Analyst . The primary responsibility of the Data Analyst in this position is to design and develop advanced artificial intelligence models using medical imaging data, with a focus on supporting clinical decision-making in emergency and trauma care. This involves conceptualizing and implementing scientifically rigorous modeling approaches, selecting and applying appropriate analytical techniques, and translating complex imaging and clinical data into actionable insights. Key responsibilities include designing algorithms and study methodologies, developing and validating statistical analysis plans, performing and interpreting analyses, and communicating results in a clear, practical manner to clinicians, researchers, and other stakeholders to improve patient outcomes in critical care settings.
Duties/Responsibilities
Due to variable nature of position, this list is to be used as a guide only.
Data Analyses, Statistical Modeling, and Machine Learning (40% of work time)
Follow detailed study protocols and analysis plans to perform a wide variety of statistical analyses, including:
Design, implement, and validate advanced deep learning models for medical imaging data, with a focus on emergency and trauma applications.
Develop and optimize 3D convolutional neural networks (CNNs), vision transformers (ViTs), and other modern architectures for clinical decision support.
Apply and evaluate a wide range of modeling techniques, including:
Image segmentation and classification.
Multimodal data integration (imaging, clinical, and demographic data).
Prediction modeling for patient outcomes and surgical decision-making.
Transfer learning and fine-tuning of pre-trained models.
Explainability and interpretability methods (e.g., saliency maps, Grad-CAM).
Conduct rigorous model validation using cross-validation, external datasets, and real-world clinical data.
Collaborate with clinicians and researchers to translate model outputs into actionable insights for use in critical care settings.
Carry out data analyses using statistical and machine learning techniques, including supervised, unsupervised, and reinforcement learning.
Develop and apply predictive models and risk scores using algorithms such as linear/logistic regression, survival analysis, PCA, GLM, GAM, GEE, SVM, random forests, neural networks, and time series models (using R or Python libraries).
Apply natural language processing (NLP) methods to unstructured text data, including bag-of-words, topic modeling (e.g., LDA), and word embeddings.
Optimize models through hyperparameter tuning and evaluate performance using appropriate metrics (e.g., RMSE, MAE, F1, AUC, sensitivity, specificity, PPV, NPV).
Prepare results for review, including visualizations, tables, and clear interpretations for reports and presentations.
Document all modeling workflows, code, and iterations with proper commenting and version control (e.g., Git).
Deploy models to production in collaboration with DevOps teams and establish systems to monitor and maintain model performance post-deployment.
Document workflows and results in a clear, reproducible, and clinically meaningful manner.
Data Exploration, Preparation, and Visualization (30% of work time)
Ensure adequate quality control by setting standards, monitor results and institute appropriate steps for data cleaning, consistency checks and other data quality control measures prior to analysis.
Maintain data documentation, physical and logical storage of scripts, records and master archive lists.
Perform descriptive and inferential descriptive analysis in R and/or Python. Write HTML/PDF/Microsoft Word reports summarizing the analysis.
Validate output tables, listings or figures generated to ensure accuracy and reliability of analyses.
Pre-process raw data to prepare for analysis. This includes cleaning and merging data from multiple sources, as well as understanding overall data quality.
Design and develop dashboard and reports with interactive visualizations using libraries such as Highcharts, Plotly, etc.
Produce high quality ad hoc and standardized reports, customized per project using R/Python procedures, tailored to different end users (e.g., clinicians, senior management, and hospital executives).
Develop efficient programs, algorithms, or systems to reduce programing time of standardized data analyses and reports.
Develop Project Analytic Plans Outlining Key Components of Analytical Approaches (15% of work time)
Work closely with colleagues and research partners to establish coding techniques, structure of dataset for studies and develop sound analytical plans.
Work with other team members to understand which analytical approach would be best suited to answer the project objectives.
Provide on-demand consultative expertise for ad-hoc requests and recommend analytical approaches to meet client needs.
Meet regularly with the project team to provide updates on project status and present results.
Write project proposals and detailed analytic plans.
Understand data requirements based on stated project goals.
Integrate with the other team members to develop multi-disciplinary analytic approaches and assist in best practices and process development.
Communication and Dissemination of Results (15% of work time)
Writing data reports containing the results of analyses.
Preparing presentations and manuscripts for different small audiences.
Providing recommendations for analytical approaches.
Providing recommendations based on results of analyses and project objectives.
Constructing elegant visualizations and dashboards to communicate findings/output.
Documenting and communicating errors in data/code to senior staff.
Performs Cross Functional and Other Duties as Assigned and/or Requested
All staff are expected to carry out their assigned duties and responsibilities in a manner which prioritizes patient and employee safety, and confidentiality. Key accountabilities in this regard include:
Strict compliance with patient/employee confidentiality practices and policies.
Strict compliance with patient/employee safety practices and standards.
Appropriate identification, reporting and response to patient/employee confidentiality breaches in accordance with established policies and procedures.
Appropriate identification, reporting and response to patient/employee safety risks and incidents/events in accordance with established policies and procedures.
Qualifications
Bachelor’s degree in mathematics, statistics, biostatistics, computer science, or related discipline with at least 1 year of professional data science experience OR demonstrable equivalent combination of specialized education and experience.
Ability to analyze and problem solve in the areas of data management and preprocessing, modeling, and evaluation, with consultation as needed.
Beginner to intermediate experience (0 to 4 years) with the following: Python, RMarkdown, Jupyter Notebooks, HTML, CSS, Javascript, Git/GitHub/GitLab.
Intermediate experience (4-5 years) with all of the following SQL, R and/or Python.
Fully proficient in the use MS Office software (Word, Excel, PowerPoint, Outlook, Internet Explorer, etc.).
Is comfortable designing and implementing medical imaging AI models using R/Python libraries under minimal supervision;
Is comfortable documenting code and using version control systems such as Git under minimal consult.
Is proficient in using common data visualization libraries from R or Python (Plotly, ggplot, matplotlib, etc.). Can build data visualizations using D3 libraries or open-source equivalents (e.g., highchart) and can prepare and automate data reports in RMarkdown or Jupyter notebooks with minimal consult.
Must be able to read in and merge data from disparate sources, perform data quality checks, manage missing data, and prepare data for machine learning models under minimal supervision.
Experience with unstructured data sources.
Experience preparing data for varied statistical methods preferred.
Experience with clinical data in a healthcare setting is a plus.
Excellent attention to detail and proven ability to learn new skills.
Excellent ability to learn new skills.
Must be able to effectively communicate with end users and managers.
Ability to analyze and problem solve in the areas of data management and preprocessing, modeling and evaluation, with consultation as needed.
Excellent organizational skills to manage multiple tasks in a timely manner, project management skills would be an asset.
Must be able to present to and train small groups.
Unity Health Toronto is committed to creating an accessible and inclusive organization. We strive to provide a recruitment process that is barrier-free and in compliance with the Accessibility for Ontarians with Disabilities Act (AODA) and the Ontario Human Rights Code. We understand that you may require an accommodation at any stage of the recruitment process. When you are contacted, please inform the Talent Acquisition Specialist and we will work with you to meet your accommodation needs. We want to emphasize that all accommodation requests are handled with the utmost confidentiality, respecting your privacy and dignity.
