While reading Chris Stolte, Diane Tang, Pat Hanrahan’s paper, I recognized the mantra “overview first, zoom and filter, then details-on-demand,” which made me feel like I was part of the data visualization’s cool kid group. The paper continues to describe two technologies they used to perform data abstraction and visual abstraction, data cubes and Polaris respectfully. Data abstraction is changing the data before mapping it to visualizations, whereas visual abstraction is changing the visual representations of the data points. Understanding dimensions (especially those greater than 3) is difficult for most people and I think the paper could have used figures/visual aides to better describe the data cubes. I did think that the figures provided after did benefit the reader but it took me a little while to parse through what was happening. The Polaris technique turns each operand into either an ordinal or a quantitative, the nominal fields turn into ordinal because they are drawn in some order! And they use the dot (.), cross (x), nest (/), and concatenate (+) operators. Then the paper describes zoom graphs, which “allow for multiple zoom paths from any given point”. A zoom can change either the visual abstraction, the data abstraction, or both. Different common zoom structures or patterns include: Chart Stacks (essentially multiple stacked line charts showing different dimensions relationship with a measure), Thematic Maps (multiple geographical levels of detail, example being a map divided by county and by state), Dependent Quantitative-Dependent Quantitative Scatterplots (like thematic maps, but the axes don’t have an inherent mapping to the real world, like a map does), and Matrices (matrix like with layers in both axes).