After identifying what a user needs to see, the next question is how they need to
see it. This doesn??™t mean how they need to view it on the screen, but how he or she
needs to analyze and slice it. A single number for the defect rate is meaningless.
What time period is covered by that number? What product does it represent? What
assembly lines produce the most or fewest defects? What shift has higher and lower
defect rates?
How an end user wants to analyze the data is a dimension. Time is the most
common dimension; almost all data can be analyzed by time. Imagine the phrase, ???I
need to see defect rates by month.??? This sentence could be rewritten as, ???I need to
see
by .??? The what portion is the fact, while the how is a dimension.
Defect rates could be viewed by month, week, day, or even hour or minute if need
be. This means that not only is time a dimension, but a dimension includes a series
of values??”in this case months, weeks, days, and so on.
Many facts are analyzed by more than one dimension. The phrase, ???I need to see
defect rates by month and assembly line??? now includes two dimensions: Time and
Assembly Line (or Location.) Imagine that this particular product is made in two
plants, each containing multiple assembly lines. It might therefore be necessary to
look at which plant has more defects, and then examine the assembly lines within
that plant. It??™s also possible to examine all the assembly lines regardless of the plant
to which they belong.
Pages:
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69