## How to create a weekly data from a monthly data?

### How to create a weekly data from a monthly data?

Hi,

MONTHLY
MARKET  PRODUCT               SPENDING  YEAR   MONTH
A       WWW                    \$10.00   2001    Jul
A       WWW                    \$10.00   2001    Aug
B       WWW                    \$30.00   2001    Jul
B       WWW                    \$40.00   2001    Aug
C       WWW                    \$10.00   2001    Jul
C       WWW                    \$20.00   2001    Aug
C       WWW                    \$30.00   2001    Sep

Basically, in the data, each product has 12 months monthly spendings within
each market. In other words,
the file has the monthly spending data by product, market, month, year.
I will need to calculate the weekly spendings and output a weekly data.

The way I calcuate the weekly spendings will be as follows:

For example,

July, 2001 - montly spending: \$100
August, 2001 - montly spending: \$200

The first 4 weeks and all 7 days were in July; therefore, the weekly
spending for the first 4 weeks will be (\$100/31)*7 because July has 31 days.
(i.e. for Sep, it will be \$100/30)

But the 5th week was different, the first two days were in July and the rest
of the 5 days were in August; therefore, the weekly spendings would be
(\$100/31)*2 + (\$200/31)*5 to get the weighted spendings.

I was wondering should I create a daily data first and convert it to a
weekly data, would it be easier to handle?  Any advice will be greatly
appreciated.

_________________________________________________________________

### How to create a weekly data from a monthly data?

Kit,
I am concerned that you are trying to get more precision out of your data
than is justified. With monthly data you have no indication of how spending was
distributed during the month. If you must average weekly spending, then I would
suggest calculating average daily spending by month and multiply that by seven.
However, be warned that trying to go from monthly data to weekly data is fraught
with dangers, particularly if there are trends and/or intra-monthly cycles (or
cycles which are not an integer multiple of months). Convert to days and then to
weeks if you must, but you are in severe danger of producing results with false
precision not justified by the input data.

Dick. March

### How to create a weekly data from a monthly data?

I concur with what* March posted on this topic.

Having said that, the easiest way to interpolate weekly observations from a
monthly time series is to use PROC EXPAND, which is in the SAS/ETS module.

Hope this helps!

Andrew

Andrew H. Karp
Sierra Information Services, Inc.
A SAS Institute Quality Partner in the USA
19229 Sonoma Highway PMB 264
Sonoma, CA 95476 USA
707/996-7380 (voice)

http://www.veryComputer.com/

I can't believe how much trouble I am having trying to convert a
weekly time series of prices to a monthly series that averages the 4
or 5 weekly prices that fall in the month.  I'd be very grateful if
anyone has an easy way to do this in 123 or excel.  I have the weekly dates
in the first column (10 years) and prices for differnt products (about 50
of them) in the next columns.  Ideally, I would end up with the monthly prices
below the columns of weekly prices (so I keep each product within one column).

Thanks,

Bill