Automating the Creation of Interactive Glyph-supplemented Scatterplots for Visualizing Algorithm Results

Ndaona is a Matlab toolkit to create interactive three-dimensional models of data often found in NLP research, such as exploring the results of classification and dimensionality reduction algorithms. Such models are useful for teaching, presentations and exploratory research (such as showing where a classification algorithm makes mistakes). Ndaona includes embedding and graphics parameter estimation algorithms, and generates files in the format of Partiview (Levy, 2001), an existing free open-source fast multidimensional data displayer that has traditionally been used in the planetarium community. Partiview1 supports a number of enhancements to regular scatterplots that allow it to display more than three dimensions’ worth of information. 1 Supplemented Scatterplots Scatterplots are not the most efficient way of representing information (Grinstein et al., 2001). However, they are intuitive and stable (Wong and Bergeron, 1997), and can be supplemented in several ways. We describe some of these augmentations in Section 1, basic Ndaona usage in Section 2, and finally a couple of some embedding methods in Section 3. http://niri.ncsa.uiuc.edu/partiview Figure 1: Regular and glyph-supplemented scatterplots showing how a linear kernel can separate happy and sad faces from the Frey Face dataset.