sravoct29 said:
Hi fakename,
Thanks for the spreadsheet..Awesome job that helps members of this forum.
If u don't mind me asking, what type of average are you performing. I mean , are u calculating arithmetic or geometric or harmonic mean etc .. or some sort of statistical median. If u are planning to add red flags into consideration when estimating time, what type of logic are u planning to use.
thanks
Hello Sravoct29,
As I mentioned the spread sheet uses averages (i.e. arithmetic mean) in the current solution. You may visit the spreadsheet that lets you see all the formulas used (by selecting a cell, and seeing the formula in the formula field at the top).
In the future solution, I am planning to use two different methods.
1. Orthodox normal distribution method (http://en.wikipedia.org/wiki/Normal_distribution) to calculate the likelihood, and the "best", and "worst" possible dates within the standard deviation.
2. Unorthodox method, where I will dynamically formulate the trend and calculate the dates.
Obviously both the calculations will use these variables
-- Total number of applications received (http://data.gc.ca/data/en/dataset/5c003561-ce36-4881-a5a4-c30f9f456690),
-- sample size (for standard deviation)
-- Cases processed (from volunteer's input)
-- New case's dates
-- And may be a couple of more which I can't think of right now
All known red flags will be entered from the volunteers, and when an applicant tries to estimate his dates only data with same level of red flags will be used, or if we don't have enough data with a flag it will estimate how a red flag affects the time period of the application (based on the limited records for that flag compared to regular records) and difference will be added to the average time.