Task: I have 57,334 claims for my firm and 997,773 claims for competitor firm. All about of laser printers.
Answer: The first idea - average count of the problems for single printer. Result 2.8 for competitor printers and 9.1 for printers in the my company.
Need to use median. For my = 1, for competitor = 2. Median is so good if we have serious outliers in the data.
Standard deviation - it's σ of mean µ. For any sets of n observation x1, x2, x3 ... xn with mean µ,
Variance = σ2 = [(x1 - µ)2 + (x2 - µ)2 + (x3 - µ)2 ... (xn - µ)2] / n
For any sets of n observation x1, x2, x3 ... xn with mean µ,
Standart deviation = σ = sqrt([(x1 - µ)2 + (x2 - µ)2 + (x3 - µ)2 ... (xn - µ)2] / n)
Percentage values can make context without it. Percentage gives the values of some figure relative to something else. Context include in the it.
Percentage points. For example. Tax is 7%. Tax change to 5% in the percentage points = 2 points. But it's tax change to 28.6% in the percentage value.
--------------------------------
Percentile - ...
--------------------------------
Conclusions
- About single variables that include many information and their essence is determine into the formula;
- Possible what we can't do single value for get any answer ...
- Possible what we have many opinions and many single value and they have various popularity. But if it's true - determining is really bad and answer will never not be truth;