DJM Topic 3 - Visualizing Proportions

Question 1

Proportion data

Answer 1

Data grouped by categories and subcategories, with the maximum, minimum, and overall distribution being important

Question 2

ranking analysis for proportions

Answer 2

Sorting the expenses to perform ranking analysis

Question 3

part-to-whole analysis for proportions

Answer 3

Expressing expenses as percentages to do part-to-whole analysis

Question 4

key questions for analyzing proportion data

Answer 4

Which department has the highest or lowest expenses? how much expenses has a specific department incurred for the whole company?

Question 5

patterns to look for in proportion data

Answer 5

Unusual differences from one value to the next, grouped values, significant breaks in a pattern, obvious exceptions to the norm

Question 6

visualizations for proportion data

Answer 6

Pie chart, donut chart, stacked area chart, stacked bar chart, treemap, pareto chart

Question 7

pie chart characteristics

Answer 7

Circle represents the whole, wedge size represents part or percentage of the whole

Question 8

when to use pie/donut charts

Answer 8

Parts make up a meaningful whole, parts are mutually exclusive, comparing parts to each other or to the whole, few parts (3 or less)

Question 9

stacked area chart use

Answer 9

Showing changes over time for several variables, percentages add up to 100% vertically, showing raw counts and peaks/valleys

Question 10

stacked bar chart characteristics

Answer 10

Similar to stacked area charts but with vertical bars, avoids angle perception problem of pie charts

Question 11

treemap characteristics

Answer 11

Rectangles of varying sizes show relative proportions, works well for hierarchical data, can use color/gradient as second measure

Question 12

pareto chart use

Answer 12

Examining cumulative contribution of parts to the whole, constructed with bar and line chart, based on pareto principle/80-20 rule

Question 13

cautions with pareto charts

Answer 13

Fluctuating systems can produce inconsistent rankings, repeated analyses needed to confirm rankings, use control chart first to ensure system stability

Question 14

when to use pareto charts

Answer 14

Data can be arranged into categories, rank of each category is important, analyzing frequency of problems/causes in a process

Question 15

techniques and best practices

Answer 15

1) Grouping categorical items ad hoc
using pareto charts with percentile
scales
2) re-expressing values for
quantitative scaling
3) using line graphs for ranking
changes over time