Data Visualization Level 2

Data Visualization Level 2

• Nonlinear Axes: In Cartesian coordinates, grid lines along an axis are evenly spaced in data units and visualization. This is called a linear scale. A nonlinear scale has uneven spacing in the visualization, even spacing in data units corresponds to uneven spacing.

• Logarithmic Scale: The most common nonlinear scale. A unit step corresponds to multiplication by a fixed value. Creating a log scale requires log transforming data values while exponentiating the numbers displayed. Numbers like 1, 3.16, 10, 31.6, and 100 placed on linear and log scales. The numbers 3.16 and 31.6 are exactly halfway between 1 and 10, and between 10 and 100, respectively.

Coordinate Systems with Curved Axes

• Polar Coordinates: Positions are specified by angle and radial distance from the origin. Useful for data with periodic values, where values at one end of the scale can be joined to the other. A practical example is the days of the year (December 31st being one day before January 1st).

Color Scales

• Use Cases: Color in data visualization has 3 fundamental use cases:

  • Distinguishing groups of data
  • Representing data values
  • Highlighting specific elements

• Qualitative Color Scales: Used to distinguish discrete items or groups without inherent order (e.g., countries on a map, manufacturers).

• Conditions of Color Scales:

  • A finite set of distinct colors, visually equivalent to each other.
  • No single color should stand out relative to others.
  • Colors should not create the impression of an order (e.g., colors getting successively lighter).

• Sequential Color Scales: Used for quantitative data values (e.g., income, temperature, speed).

  • Colors should clearly indicate larger/smaller values and the distance between values.
  • The color scale should appear to vary uniformly across its entire range. Examples include a single hue shift or multiple hues (e.g., red to light yellow).

• Diverging Color Scales: Used to visualize the deviation of data values relative to a neutral midpoint. Useful for data with positive and negative values. The progression from light to dark colors on either side of the midpoint should be balanced.

• Highlighting with Color: Colors can be used to emphasize specific categories or values in a dataset that carry key information. This helps strengthen a story.

Directory of Visualizations

• Amounts: The most common way to visualize amounts is using bars (vertical or horizontal).
  • Alternatives include dots placed where bars would end if there are more than 1 set of categories.
  • Categories can be grouped or stacked
  • Using a heatmap, with categories on x and y axes and amount shown by a color.

Distributions

• Visualizing Distributions: Histograms and density plots visually represent distributions but require arbitrary parameter choices. They may be misleading.

  • Cumulative density and quantile-quantile plots accurately represent data but may be harder to interpret.

• Other Useful Plots:

  • Boxplots, violin plots, strip charts, and sina plots are useful to visualize many distributions at once.

Proportions

• Visualizing Proportions: Pie charts, side-by-side/stacked bars are ways to visualize proportions. Pie charts emphasize parts that add up to a whole and simplify fractions.
  • Side-by-side bars allow comparison of individual parts. Stacked bars are for multiple sets of proportions
  • Stacked density plots are useful if proportions change along a continuous variable.
  • Mosaic plots, treemaps, and parallel sets are useful when visualizing proportions based on multiple variables. Mosaic plots combine every level of variable with every possible level of another variable. Treemaps don't require this. Parallel sets are helpful when more than 2 variables are being compared.

x-y Relationships

• Scatterplots: Shows the relationship between two quantitative variables. Dot size can be used to represent a third quantitative variable (creating a bubble chart).
  • Paired data can be displayed with a line indicating x = y, or a slopegraph.
  • For large datasets, scatterplots can be uninformative due to overplotting. Contour lines, 2D bins, or hex bins may be helpful instead
  • If visualizing more than two quantities, correlation coefficients may be shown as a correlogram instead of the raw data.

This quiz explores advanced topics in data visualization, focusing on nonlinear axes and logarithmic scales. Additionally, it covers coordinate systems with curved axes, such as polar coordinates, and their applications in visualizing periodic data. Test your understanding of complex visualization techniques.