Introduction to Data Management and Analytics for Business
Agenda
Now that we know how to clean and transform data, we can work toward describing/representing our data in a dashboard for decision-makers.
This week, we'll be covering several analyses/visualizations of considerable relevance to a decision support dashboard:
Multidimensional analysis (continuing the Lending Club case)
What-if analysis
Trend analysis
Z-score analysis
Benford analysis
***First, we will also clarify the difference between computing measures and columns
Average of averages (can happen with columns)
The average of averages problem exists when groups have different N. It biases the statistics because the N used for the average is not the population’s N.
Class A (student scores): 9, 9, 9, 9, 9, 9, 9, 9, 9, 9 10 students with high scores Average: (9+9+9+9+9+9+9+9+9+9)/10 = 9
Class B (student scores): 9, 1 1 student with high score and 1 student with low score Average: (9+1)/2 = 5
Population’s average: (9+9+9+9+9+9+9+9+9+9+9+1)/12 =
Average of averages: (9+5)/2 =
Because the Ns of the two classes are different and we are giving equal weight two the two groups, the average of the averages is not representative of the population’s average.