Projects

Lossless Visualization of 4D Compositional Data on a 2D Canvas

arXiv Preprint

The paper suggests the method to visualize 4D compositional data on a 2D canvas. This can be used to visualize data such as posterior model probabilities for 4 models on (2D) paper without loss of information and distortion. My contributions are general polishing of manuscript, creating 3D interactive figures and animation, and development of R package (planned).

---

R package “metabmc”

The package implements meta-uncertainty quantification for Bayesian model comparison. This can be used to evaluate trustworthiness of posterior model probabilities, which tend to take extreme values. Check the project website for more detail. The development is still in progress and will be published in near future.

---

R package for Bayesian proxy structural VAR

The package implements Bayesian vector autoregressive model with proxy variables for structural shocks in economy. This can help agents like government and central bank to make decisions, considering the expected effects in the system of economy. The development is still in progress and will be published in near future.