This repository contains the data and R code for the manuscript: Data Driven Methodology COVID19 Related Pharmacovigilance

The R code consists of five parts:
1. ebgm: calculate N, E, RR, PRR, EBGM and CI of EBGM for the drug-adverse event pairs regarding large number of spontaneouse reports.
2. glasso_circosplot: get a sparse inverse covariance matrix using L1 penalty based on data from ebgm, and plot the sparse matrix in a circos layout with edgebundling for better visualizaiton.
3. pca: 2D and 3D visualizing drugs based on the variation of pulmonary ADE.
4. friedman_test: nonlinear multiple comparsions for drugs in different grouping schemes, drug features consists of pulmonray ADE features.
5. arcdiagram: using drug-drug and drug-ADE covariance matrice to construct a more comprehensive diagram to visualize and compare drugs based on their pulmonary ADEs.

