NXG Logic will soon roll out a new workflow-based version of Explorer, which will provide for multiple analyses in one run. Collaboration with colleagues will also become more streamlined as data and workflows can be shared, opened by collaborators, and run the exact same way the original workflow was run. In this way, discussion surrounding each module that is processed won't need to be partitioned over a slide deck for presentation during meetings and conference calls. The entire workflow can be run by a collaborator/customer who may have questions for various assumptions and/or output or walked through by the analyst directly.
Pre-canned templates for various analyses have also been generated, which will result in a significant time-savings, as a full pipeline run, and data processed with outputs will be centrally located while using real-time results. Therefore, you will no longer need to make a Powerpoint or Word to go through results with customers, just ask them to run the workflow using their own copy of Explorer. It will also enable more productivity for lecturers and instructors, by making computer labs more interesting with minimal misunderstanding and misinterpretation.
In the workflows shown below, at the top is shown the icons for summary statistics (statistics, category frequency tables, and automatically generated histograms for all features) and correlation, which can run correlation, covariance, and distance matrices for Euclidean, Manhattan, Chebyshev, Canberra, Tanimoto, etc. Below association is an ANOVA run, which simultaneously runs Kruskal-Wallis as well as Bartlett's test for heteroscedacticity. This is followed by a paired t-test (with Wilcoxon signed ranked), multiple linear regression, and multivariate regression. Unsupervised class discovery analysis includes a variety of non-linear (linear) manifold learning methods, which is followed by class prediction with a variety of classifiers providing outputs for k-fold cross-validation, sens/spec, ROC-AUC, and ensemble classifier diversity.
