theory lecture 7 Flashcards
Ancombe’s quartet
shows four different graphs that have the same values for a lot of characteristics, eg. number of points, average x, etc. they look very different when visualised.
general visualisation principles
- Show the data: grids, tick marks, and explanatory texts should be avoided.
- Simplify: choose the graphic that most effectively communicates the information.
aspect ratio
if you compress or expand your graph it will take on a different shape.
super tables
tables that show multiple dimensions within your data.
scatterplot matrix
shows a scatterplot for each pair of dimensions and the relations between them. all combinations are arranged in a grid or matrix with a row or column for each dimension. it looks for clusters, outliers, partial correlations, and trends.
trellis plot
parallel coordinates
each variable/dimension is a vertical line. the bottom of the line indicates a low value, the top a high value. each record creates a polyline across all dimensions. similar records cluster on the screen. it looks for clusters, outliers, line angles, and crossings.
glyphs
shapes whose attributes are controlled by data values.
star glyphs
a set of N rays spaced at equal angles. the length of each ray is proportional to the value fot that dimension. the line connects all endpoints of a shape. you can look for shape similarities and differences, and trends.
dimensional stacking
breaks each dimension range into bins. it breaks the screen into a grid using the number of bins for 2 dimensions. the process is repeated for 2 more dimensions within the subimages formed by the first grid and recurses through all dimensions. it looks for repeated patterns, outliers, trends, and gaps. it is an interactive visualisation.
drilling down
zooming into data to get more finely grained data.
rolling up
zooming out of your data.
Chord diagram
a way of data clustering.
time series
if data changes over time.
