Biostatistics Weekly Seminar

Data analyses have grown more complex over time in part due to tremendous advances in data collection, measurement technology, and computational power. These advances have allowed us to measure the world in ever greater detail and apply complex models from which we can learn about the underlying phenomena being studied. However, the unrestricted and undisciplined analyses of complex datasets has lead to a proliferation of non-reproducible findings. As the data science revolution continues forward and touches all areas of society, we propose that there is a need to specify the craft of data analysis in more formal terms. Some benefits of formalizing the data analysis process include the development of novel metrics of data analysis quality, the articulation of principles for analytic design, and the ability to scale the teaching of data analysis to large audiences. We will present a theoretical framework for the analytic process and describe its potential for improving the quality of data analysis.

Virtual: Zoom Link to Follow
1 November 2023
1:30pm