Both visual analytics tracks use the same statistical methods. For the correlation network, OmicsAnalyst uses the loading scores from dimension reduction analysis to select the top features to construct network; while the dimension reduction results are directly explored within 3D scatter plots (scores, loadings, biplots).
The two tools offer complementary information on the data. We suggest first performing dimension reduction with the 3D scatter plot to see how well the reduced dimensions separate interesting groups of samples. Then, these reduced dimensions can be examined in more detail using the network to see which features they are composed of, and how these features relate to each other.