Given the current raw Json output of the type and stats mode, it's cleaner to display them as flatten table view. Each row is a field (the nested field uses dot notation). And we can merge both type and stats modes into one table, and highlight those problematic fields (e.g., multiple types).