Data Intelligence, Business Analytics
I know that for a monthly series of data, there are 2 well known programs that are free and reliable enough to use:
2) X-12 ARIMA Graph
Does anyone have any recommendations on programs / techniques to obtain Seasonally adjusted annualized sales (for New Homes) derived from a long history of weekly sales (since 1988)?
Some challenges include:
The weeks sometimes span across 2 different months. Also, Holidays will occur in different weeks depending on the year.
I have access to resale closings on a daily basis, and that data may be useful to draw conclusions about "day of the week" information (6 years of history).
Any thoughts would be much appreciated!
At this time I believe I will evaluate the resale sales that occur during any 7 day period that spans 2 different months to see the % distribution of weekly sales (since I can see daily Resale totals). Then, I will assign the "New Home" weekly sales by the same % distribution as the "Resale Homes". Then with my monthly numbers in place, I can use X-12 and determine the months SAAS rates.
I believe Resale homes have very similar demand components (and less random supply constraints), so it might be fitting to combine all sales and derive the SAAS from that number. However, if I was to use the method described above, would that pollute the combined data set?