L4: Data visualization Flashcards
EDA
Understand your data
Data types
Discrete vs. continous
Orded vs. unordered
Orded x discret: Ordinal, quantitative
Unordered x discrete: Nominal, categories
Continous x ordered: Fields
Continous x unordered: Cyclic values
Discrete vs…
continious
Ordered data vs.
unordered data
Is ordinal (eg. size S, M, L, XL…) discrete or continuous?
Discrete
Is Fields (eg. altitude, temperature) discrete or continuous?C
Continuous
Is nominal (eg. shape) discrete or continuous=
Discrete
Is Cyclic values (eg. directions, hues) discrete or continuous
Continuous
Is categories (eg. nationality) discrete or continuous?
Discrete
Is quantitative (1,2,3…) discrete or continuous?
Discrete
Science of graph
Position, length, angle/slope, area, volume, color/density
The principle of proportional ink?
When a shaded region is used to represent a numerical value, the area of that shaded region should be directly proportional to the corresponding value
Causes for bad graph
Bad taste, data and perception
Three questions to ask when creating a good graph
- What is your point (headline)
- How can your emphasize your point in your chart (chart type, data, color, highlights, annotations)
What does the final chart show excatly? (description, legend/keys, source)
Principles for good graph
Principles for good graphs
· Choose the right type of plot (based on the number of variables, type)
· Pay attention to the match between variable type and graphical element
· Be mindful when choosing colors
· Label your axes
· Always start axes at 0 (unless you don’t, but then be clear about it)
· Make sure people can read the axes
Keep it simple