I want to fit a skew normal curve to a data set in excel. My goal is to model incomes in geographic areas in order to get the # of people < $23,000 income, using several median values and their associated population sizes. I graphed the population frequency distributions in excel - some are normal as expected, but some appear skewed. Ive had success modelling with the cumulative normdist function in excel, as long as the data's frequency distribution was normal-looking and not skewed.
To get the skew pdf, do I have to calculate the error function, or just use the cdf of standard normal times the st norm pdf (times 2)?
Is there a way to integrate the skew normal pdf (once obtained) in excel perhaps using rectangles under the fitted curve? How to test goodness-of-fit?
Thanks for your advice.