Overview and Background
Bayesian statistical models can be used to analyze data sets. Compared to the frequentist approach they include some interesting features like the explicit statement of priors. As any statistical model they are based on several assumptions that are not trivial to be checked automatically. Statisticians often use visualizations to help them inspecting their models. However, they usually use unconnected static visualizations.
The main question to be elaborated in this project is whether an interactive visual analytics approach to some class of Bayesian models can add additional value beyond the state-of-the-art approach. The goal is to compose an integrated application that guides analysts through the workflow, shows suitable visualizations when needed, helps in specifying parameters and aids in assessing the estimation process
as well as its results. Ideally, the system automatically points out problems and suggests potential solutions. To begin with, the system shall be focused on a particular problem set and show that a visual analytics approach can help analysts do their work.
- Gabry, J., Simpson, D., Vehtari, A., Betancourt, M., and Gelman, A. (2018). Visualization in Bayesian workflow https://arxiv.org/abs/1709.01449
- Gelman, A., J. B. Carlin, H. S. Stern, D. B. Dunson, A. Vehtari, and D. B. Rubin (2013). Bayesian Data Analysis (Third ed.). Chapman & Hall/CRC. http://www.stat.columbia.edu/~gelman/book/